
%reset -f

# jupyter_pwd = %pwd
# if jupyter_pwd == "/":
#     %cd /workspace


# ipynb形式のライブラリのインポート
%run ./lib/lib.ipynb

# 生データの入ったCSVファイルの保持されたディレクトリ名を格納している変数
csvDirPath = "./csv_files/"

# NPBのベンチマーク名のリスト
benchmarkNames = ["cg", "ep", "ft", "is", "lu", "mg"]

# LULESH ベンチマークプログラムのプロセス数・問題サイズ・イテレーション数
lulesh_processes: list[int] = [8, 27, 64, 125, 216, 343, 512]
lulesh_iterations: list[int] = [8, 16, 32, 64, 128, 256]
lulesh_sizes: list[int] = [16, 24, 32, 48, 64, 128]

# ipynb形式のライブラリノートを.py形式に変更したものをインポート
import lib
import lib.lab_lib
from lib.lab_lib import *

train_lulesh_processes: list[int] = [8, 27, 64, 125, 216, 343]
train_lulesh_iterations: list[int] = [8, 16, 32, 64, 128]
train_lulesh_sizes: list[int] = [16, 24, 32, 48]

test_lulesh_processes: list[int] = [512, 729, 1000]
test_lulesh_iterations: list[int] = [256, 512, 1024]
test_lulesh_sizes: list[int] = [64, 96, 128]

list_modelName: list[str] = [
    "modelIp",
    "modelLog",
    "modelLinAndIp",
    "modelLinAndLog",
    "modelIpAndLin",
    "modelIpAndLog",
    "modelLogAndLin",
    "modelLogAndIp",
    "modelProcessDividedByProblemSize",
    "modelProblemSizeDividedByProcess",
    "modelInfiniteProductOfProblemSizeMultipliedByProcesses",
    "modelInfiniteProductOfProblemSizeDividedByProcesses",
    "modelLinearSumOf2elementCombination",
    "modelLinearSumOfElementCombinations",
    "modelLinearSumOf2elementCombinationWithSquared",
    "modelLinearSumOf2elementCombinationWithCubed",
    "modelSquareRootOfProcess",
    "modelSquareRootTimesOtherElems",
    "modelObeyOneParameter",
    "modelLin"
    # "modelBasicTree",
]


# ipynb形式のライブラリのインポート
%run ./lib/lib.ipynb

# ipynb形式のライブラリノートを.py形式に変更したものをインポート
import lib
import lib.lab_lib
from lib.lab_lib import *

# 生データの入ったCSVファイルの保持されたディレクトリ名を格納している変数
csvDirPath = "./csv_files/"

# NPBのベンチマーク名のリスト
benchmarkNames = ["cg", "ep", "ft", "is", "lu", "mg"]

# NPBのプロセス数
npb_process: list[int] = [2, 4, 8, 16, 32, 64, 128, 256]
train_npb_process: list[int] = npb_process[:-1]
test_npb_process: list[int] = npb_process[-1:]
# NPBのCGの初期変数
cg_na: list[int] = [14000, 30000, 75000, 100000, 1500000]
cg_nonzer: list[int] = [11, 12, 13, 14, 15, 18, 21]
cg_niter: list[int] = [15, 30, 75, 90, 100]
cg_shift: list[int] = [20, 40, 60, 80, 110, 200]

train_cg_na: list[int] = cg_na[:-1]
train_cg_nonzer: list[int] = cg_nonzer[:-1]
train_cg_niter: list[int] = cg_niter[:-1]
train_cg_shift: list[int] = cg_shift[:-1]

test_cg_na: list[int] = cg_na[-1:]
test_cg_nonzer: list[int] = cg_nonzer[-1:]
test_cg_niter: list[int] = cg_niter[-1:]
test_cg_shift: list[int] = cg_shift[-1:]

# LULESH ベンチマークプログラムのプロセス数・問題サイズ・イテレーション数
lulesh_processes: list[int] = [8, 27, 64, 125, 216, 343, 512]
lulesh_iterations: list[int] = [8, 16, 32, 64, 128, 256]
lulesh_sizes: list[int] = [16, 24, 32, 48, 64, 128]

train_lulesh_processes: list[int] = [8, 27, 64, 125, 216, 343]
train_lulesh_iterations: list[int] = [8, 16, 32, 64, 128]
train_lulesh_sizes: list[int] = [16, 24, 32, 48]

test_lulesh_processes: list[int] = [512, 729, 1000]
test_lulesh_iterations: list[int] = [256, 512, 1024]
test_lulesh_sizes: list[int] = [64, 96, 128]

# Extra-Pのオプション
modelerNames: list[str] = [
    # "refining",
    "multi-parameter",
    "default",
    # "basic --options poly_exponents=-1,0,1,2,3 log_exponents=0,1 force_combination_exponents=1 allow_negative_exponents=1"
]

modelerOption: str = """ --options \#spm=Basic \#spo=poly_exponents=-1,0,1,2,3,log_exponents=0,1,force_combination_exponents=1,allow_negative_exponents=True"""

list_modelName: list[str] = [
    "modelIp",
    "modelLog",
    "modelLinAndIp",
    "modelLinAndLog",
    "modelIpAndLin",
    "modelIpAndLog",
    "modelLogAndLin",
    "modelLogAndIp",
    "modelProcessDividedByProblemSize",
    "modelProblemSizeDividedByProcess",
    "modelInfiniteProductOfProblemSizeMultipliedByProcesses",
    "modelInfiniteProductOfProblemSizeDividedByProcesses",
    "modelLinearSumOf2elementCombination",
    "modelLinearSumOfElementCombinations",
    "modelLinearSumOf2elementCombinationWithSquared",
    "modelLinearSumOf2elementCombinationWithCubed",
    "modelSquareRootOfProcess",
    "modelSquareRootTimesOtherElems",
    "modelObeyOneParameter",
    "modelLin"
    # "modelBasicTree",
]
list_csvDir = [
    "./csv_files/lulesh_1st/",
    "./csv_files/lulesh_2nd/",
    "./csv_files/lulesh_3rd/",
]

DEBUG:__main__:hello
DEBUG:lib.lab_lib:hello
DEBUG:__main__:hello


date: str = "2022年10月10日"

input_rawDF: pd.DataFrame = return_rawDFinLULESH(
    processes=lulesh_processes,
    iterations=lulesh_iterations,
    sizes=lulesh_sizes,
    csvDirPath=csvDirPath,
)

input_rawDF = input_rawDF.rename(columns={"Name": "functionName"})

input_list_expVarNames: list[str] = ["process", "iteration", "size"]
input_list_resVarNames: list[str] = ["#Call"]


list_modelName_before: list[str] = [
    "modelIp",
    "modelLog",
    "modelLinAndIp",
    "modelLinAndLog",
    "modelIpAndLin",
    "modelIpAndLog",
    "modelLogAndLin",
    "modelLogAndIp",
    "modelProcessDividedByProblemSize",
    "modelProblemSizeDividedByProcess",
    # "modelInfiniteProductOfProblemSizeMultipliedByProcesses",
    # "modelInfiniteProductOfProblemSizeDividedByProcesses",
    "modelLinearSumOf2elementCombination",
    "modelLinearSumOfElementCombinations",
    # "modelLinearSumOf2elementCombinationWithSquared",
    # "modelLinearSumOf2elementCombinationWithCubed",
    # "modelSquareRootOfProcess",
    # "modelSquareRootTimesOtherElems",
    # "modelObeyOneParameter",
    "modelLin"
    # "modelBasicTree",
]

# 集合演算を用いて説明変数・目的変数が入力DFに存在していることを確認
if input_list_expVarNames == [] or (
    set(input_list_expVarNames).issubset(set(input_rawDF.columns.tolist())) == False
):
    warnings.warn(f"説明変数の指定に誤りがあります")
    # return False
if input_list_resVarNames == [] or (set(input_rawDF.columns.tolist()) == False):
    warnings.warn(f"問題変数の指定に誤りがあります")
    # return False

function_names: list[str] = list(set(input_rawDF["functionName"].tolist()))
if len(function_names) == 0:
    warnings.warn("与えられたDFに関数名がありません")
    # return False

result_series_list: list[pd.Series] = []

for function_name in function_names:
    # 関数ごとの生データ
    input_rawDF_per_function = input_rawDF[input_rawDF["functionName"] == function_name]
    # モデルの構築
    models = Models(
        inputDF=input_rawDF_per_function,
        expVarColNames=input_list_expVarNames,
        resVarColNames=input_list_resVarNames,
        targetDF=None,
        modelNames=list_modelName_before,
    )

    with warnings.catch_warnings():
        warnings.simplefilter("ignore")
        models.setUpDataBeforeCalcLr()
        models.calcLr()
        models.calcMAPE()

    dictCalcedMAPE = models.returnCalculatedMAPE()

    # 算出されたMAPEの数値をfloatにする
    for key in dictCalcedMAPE.keys():
        dictCalcedMAPE[key] = float(dictCalcedMAPE[key])

    # 関数ごとの結果に格納
    dict_for_series: dict = copy.deepcopy(dictCalcedMAPE)
    dict_for_series["functionName"] = function_name

    series = pd.Series(dict_for_series)
    result_series_list.append(series)

resultDF: pd.DataFrame = pd.DataFrame(result_series_list)
resultDF = addLowestMAPEColumn(
    inputDF=resultDF, model_name_list=list_modelName_before, version=2
)
resultDF.mean()

/tmp/ipykernel_121/3224927892.py:92: FutureWarning: The default value of numeric_only in DataFrame.mean is deprecated. In a future version, it will default to False. In addition, specifying 'numeric_only=None' is deprecated. Select only valid columns or specify the value of numeric_only to silence this warning.
  resultDF.mean()

modelLin                               324.173667
modelIp                                411.497704
modelLog                               454.526334
modelProcessDividedByProblemSize        40.081234
modelProblemSizeDividedByProcess        35.645728
modelLinAndIp                          421.093841
modelLinAndLog                         455.735629
modelIpAndLin                          315.660772
modelIpAndLog                          446.710080
modelLogAndLin                         323.795580
modelLogAndIp                          419.026472
modelLinearSumOf2elementCombination    194.468150
modelLinearSumOfElementCombinations     41.843775
最低値                                      4.756974
dtype: float64


date: str = "2022年10月10日"

input_rawDF: pd.DataFrame = return_rawDFinLULESH(
    processes=train_lulesh_processes,
    iterations=train_lulesh_iterations,
    sizes=train_lulesh_sizes,
    csvDirPath=csvDirPath,
)

input_rawDF = input_rawDF.rename(columns={"Name": "functionName"})

input_list_expVarNames: list[str] = ["process", "iteration", "size"]
input_list_resVarNames: list[str] = ["#Call"]

list_modelName: list[str] = [
    "modelIp",
    "modelLog",
    "modelLinAndIp",
    "modelLinAndLog",
    "modelIpAndLin",
    "modelIpAndLog",
    "modelLogAndLin",
    "modelLogAndIp",
    "modelProcessDividedByProblemSize",
    "modelProblemSizeDividedByProcess",
    "modelInfiniteProductOfProblemSizeMultipliedByProcesses",
    "modelInfiniteProductOfProblemSizeDividedByProcesses",
    "modelLinearSumOf2elementCombination",
    "modelLinearSumOfElementCombinations",
    "modelLinearSumOf2elementCombinationWithSquared",
    "modelLinearSumOf2elementCombinationWithCubed",
    "modelSquareRootOfProcess",
    "modelSquareRootTimesOtherElems",
    "modelObeyOneParameter",
    "modelLin"
    # "modelBasicTree",
]

# 集合演算を用いて説明変数・目的変数が入力DFに存在していることを確認
if input_list_expVarNames == [] or (
    set(input_list_expVarNames).issubset(set(input_rawDF.columns.tolist())) == False
):
    warnings.warn(f"説明変数の指定に誤りがあります")
    # return False
if input_list_resVarNames == [] or (set(input_rawDF.columns.tolist()) == False):
    warnings.warn(f"問題変数の指定に誤りがあります")
    # return False

function_names: list[str] = list(set(input_rawDF["functionName"].tolist()))
if len(function_names) == 0:
    warnings.warn("与えられたDFに関数名がありません")
    # return False

result_series_list: list[pd.Series] = []

for function_name in function_names:
    # 関数ごとの生データ
    input_rawDF_per_function = input_rawDF[input_rawDF["functionName"] == function_name]
    # モデルの構築
    models = Models(
        inputDF=input_rawDF_per_function,
        expVarColNames=input_list_expVarNames,
        resVarColNames=input_list_resVarNames,
        targetDF=None,
        modelNames=list_modelName,
    )

    with warnings.catch_warnings():
        warnings.simplefilter("ignore")
        models.setUpDataBeforeCalcLr()
        models.calcLr()
        models.calcMAPE()

    dictCalcedMAPE = models.returnCalculatedMAPE()

    # 算出されたMAPEの数値をfloatにする
    for key in dictCalcedMAPE.keys():
        dictCalcedMAPE[key] = float(dictCalcedMAPE[key])

    # 関数ごとの結果に格納
    dict_for_series: dict = copy.deepcopy(dictCalcedMAPE)
    dict_for_series["functionName"] = function_name

    series = pd.Series(dict_for_series)
    result_series_list.append(series)

resultDF: pd.DataFrame = pd.DataFrame(result_series_list)
resultDF = addLowestMAPEColumn(
    inputDF=resultDF, model_name_list=list_modelName, version=2
)
resultDF.mean()

/tmp/ipykernel_121/3818556874.py:91: FutureWarning: The default value of numeric_only in DataFrame.mean is deprecated. In a future version, it will default to False. In addition, specifying 'numeric_only=None' is deprecated. Select only valid columns or specify the value of numeric_only to silence this warning.
  resultDF.mean()

modelLin                                                  29.441654
modelIp                                                   59.154925
modelLog                                                  52.719367
modelProcessDividedByProblemSize                          29.798790
modelProblemSizeDividedByProcess                          29.624622
modelInfiniteProductOfProblemSizeDividedByProcesses       32.285411
modelInfiniteProductOfProblemSizeMultipliedByProcesses    35.270480
modelLinAndIp                                             66.857906
modelLinAndLog                                            54.345673
modelIpAndLin                                             22.268142
modelIpAndLog                                             47.307227
modelLogAndLin                                            28.010418
modelLogAndIp                                             64.735627
modelLinearSumOf2elementCombination                       34.761001
modelLinearSumOfElementCombinations                        3.261087
modelLinearSumOf2elementCombinationWithSquared            38.297241
modelLinearSumOf2elementCombinationWithCubed              40.734855
modelSquareRootOfProcess                                  28.765539
modelSquareRootTimesOtherElems                            64.499527
modelObeyOneParameter                                     29.441654
最低値                                                        0.475034
dtype: float64


# 2022年10月14日 研究室内発表用 資料作成

# |<既存モデル>|<すべてのモデル>|
# |<MAPE@既存モデル>|<MAPE＠すべてのモデル>|

expVar: list[str] = ["process", "iteration", "size"]
resVar: str = "#Call"

trainDF: pd.DataFrame = ret_averaged_rawDF_lulesh(
    list_process=train_lulesh_processes,
    list_iteration=train_lulesh_iterations,
    list_size=train_lulesh_sizes,
    list_csvDir=list_csvDir,
    resVar=resVar,
)

trainDF = trainDF.reset_index()

trainDF["functionName"] = trainDF["Name"]

functionNames: list[str] = sorted(list(set(trainDF["Name"])))

_result_series_list: list[pd.Series] = []
_result_series_list_before: list[pd.Series] = []
for functionName in functionNames:
    trainDF_perFunc: pd.DataFrame = trainDF[trainDF["functionName"] == functionName]

    models = Models(
        inputDF=trainDF_perFunc,
        expVarColNames=expVar,
        resVarColNames=[resVar],
        targetDF=None,
        modelNames=list_modelName,
    )

    models_before = Models(
        inputDF=trainDF_perFunc,
        expVarColNames=expVar,
        resVarColNames=[resVar],
        targetDF=None,
        modelNames=list_modelName_before,
    )

    models.setUpDataBeforeCalcLr()
    models.calcLr()
    models.calcMAPE()

    models_before.setUpDataBeforeCalcLr()
    models_before.calcLr()
    models_before.calcMAPE()

    dictCalcedMAPE = models.returnCalculatedMAPE()
    dictCalcedMAPE_before = models_before.returnCalculatedMAPE()

    for key in dictCalcedMAPE.keys():
        dictCalcedMAPE[key] = float(dictCalcedMAPE[key])
    for key in dictCalcedMAPE_before.keys():
        dictCalcedMAPE_before[key] = float(dictCalcedMAPE_before[key])

    dict_for_series: dict = copy.deepcopy(dictCalcedMAPE)
    dict_for_series["functionName"] = functionName

    dict_for_series_before: dict = copy.deepcopy(dictCalcedMAPE_before)
    dict_for_series_before["functionName"] = functionName

    series: pd.DataFrame = pd.Series(dict_for_series)
    series_before: pd.DataFrame = pd.Series(dict_for_series_before)

    _result_series_list.append(series)
    _result_series_list_before.append(series_before)

resultDF: pd.DataFrame = pd.DataFrame(_result_series_list)
resultDF = addLowestMAPEColumn(
    inputDF=resultDF, model_name_list=list_modelName, version=2
)
resultDF_before: pd.DataFrame = pd.DataFrame(_result_series_list_before)
resultDF_before = addLowestMAPEColumn(
    inputDF=resultDF_before, model_name_list=list_modelName_before, version=2
)

/usr/local/lib/python3.10/site-packages/scipy/optimize/_minpack_py.py:881: OptimizeWarning: Covariance of the parameters could not be estimated
  warnings.warn('Covariance of the parameters could not be estimated',
/usr/local/lib/python3.10/site-packages/scipy/optimize/_minpack_py.py:881: OptimizeWarning: Covariance of the parameters could not be estimated
  warnings.warn('Covariance of the parameters could not be estimated',
/usr/local/lib/python3.10/site-packages/scipy/optimize/_minpack_py.py:881: OptimizeWarning: Covariance of the parameters could not be estimated
  warnings.warn('Covariance of the parameters could not be estimated',
/usr/local/lib/python3.10/site-packages/scipy/optimize/_minpack_py.py:881: OptimizeWarning: Covariance of the parameters could not be estimated
  warnings.warn('Covariance of the parameters could not be estimated',
/usr/local/lib/python3.10/site-packages/scipy/optimize/_minpack_py.py:881: OptimizeWarning: Covariance of the parameters could not be estimated
  warnings.warn('Covariance of the parameters could not be estimated',
/usr/local/lib/python3.10/site-packages/scipy/optimize/_minpack_py.py:881: OptimizeWarning: Covariance of the parameters could not be estimated
  warnings.warn('Covariance of the parameters could not be estimated',
/usr/local/lib/python3.10/site-packages/scipy/optimize/_minpack_py.py:881: OptimizeWarning: Covariance of the parameters could not be estimated
  warnings.warn('Covariance of the parameters could not be estimated',
/usr/local/lib/python3.10/site-packages/scipy/optimize/_minpack_py.py:881: OptimizeWarning: Covariance of the parameters could not be estimated
  warnings.warn('Covariance of the parameters could not be estimated',
/usr/local/lib/python3.10/site-packages/scipy/optimize/_minpack_py.py:881: OptimizeWarning: Covariance of the parameters could not be estimated
  warnings.warn('Covariance of the parameters could not be estimated',
/usr/local/lib/python3.10/site-packages/scipy/optimize/_minpack_py.py:881: OptimizeWarning: Covariance of the parameters could not be estimated
  warnings.warn('Covariance of the parameters could not be estimated',


# resultDF


# resultDF_before


print(resultDF.mean())
print(resultDF_before.mean())

modelLin                                                  222179.489821
modelIp                                                   223497.366933
modelLog                                                  223015.868812
modelProcessDividedByProblemSize                             114.335746
modelProblemSizeDividedByProcess                             109.622295
modelInfiniteProductOfProblemSizeDividedByProcesses          112.047111
modelInfiniteProductOfProblemSizeMultipliedByProcesses       112.634694
modelLinAndIp                                             224937.799479
modelLinAndLog                                            225657.851702
modelIpAndLin                                             224677.364035
modelIpAndLog                                             228642.094382
modelLogAndLin                                            217973.966443
modelLogAndIp                                             219952.365344
modelLinearSumOf2elementCombination                       185327.084954
modelLinearSumOfElementCombinations                       229591.660065
modelLinearSumOf2elementCombinationWithSquared            232451.512990
modelLinearSumOf2elementCombinationWithCubed              225786.540482
modelSquareRootOfProcess                                  217637.402990
modelSquareRootTimesOtherElems                            192621.816842
modelObeyOneParameter                                     222179.489821
最低値                                                           96.722879
dtype: float64
modelLin                               222179.489821
modelIp                                223497.366933
modelLog                               223015.868812
modelProcessDividedByProblemSize          114.335746
modelProblemSizeDividedByProcess          109.622295
modelLinAndIp                          224937.799479
modelLinAndLog                         225657.851702
modelIpAndLin                          224677.364035
modelIpAndLog                          228642.094382
modelLogAndLin                         217973.966443
modelLogAndIp                          219952.365344
modelLinearSumOf2elementCombination    185327.084954
modelLinearSumOfElementCombinations    229591.660065
最低値                                        98.068361
dtype: float64

/tmp/ipykernel_121/587506451.py:1: FutureWarning: The default value of numeric_only in DataFrame.mean is deprecated. In a future version, it will default to False. In addition, specifying 'numeric_only=None' is deprecated. Select only valid columns or specify the value of numeric_only to silence this warning.
  print(resultDF.mean())
/tmp/ipykernel_121/587506451.py:2: FutureWarning: The default value of numeric_only in DataFrame.mean is deprecated. In a future version, it will default to False. In addition, specifying 'numeric_only=None' is deprecated. Select only valid columns or specify the value of numeric_only to silence this warning.
  print(resultDF_before.mean())


# CG


trainDF_CG: pd.DataFrame = return_rawDF_cg(
    list_process=train_npb_process,
    list_na=train_cg_na,
    list_nonzer=train_cg_nonzer,
    list_niter=train_cg_niter,
    list_shift=train_cg_shift,
    csvDir=csvDirPath,
)

testDF_CG: pd.DataFrame = return_rawDF_cg(
    list_process=test_npb_process,
    list_na=test_cg_na,
    list_nonzer=test_cg_nonzer,
    list_niter=test_cg_niter,
    list_shift=test_cg_shift,
    csvDir=csvDirPath,
)

trainDF_CG = trainDF_CG.reset_index()
testDF_CG = testDF_CG.reset_index()

trainDF_CG["functionName"] = trainDF_CG["Name"]
testDF_CG["functionName"] = testDF_CG["Name"]

functionNames_CG: list[str] = sorted(list(set(trainDF_CG["Name"])))

expVar_CG: list[str] = ["process", "na", "nonzer", "niter", "shift"]
resVar_CG: str = "#Call"

dict_symbols_CG = {}
for elem in expVar_CG:
    dict_symbols_CG[elem] = symbols(elem, real=True)

target_env_cg = [
    (dict_symbols_CG["process"], test_npb_process[-1]),
    (dict_symbols_CG["na"], test_cg_na[-1]),
    (dict_symbols_CG["nonzer"], test_cg_nonzer[-1]),
    (dict_symbols_CG["niter"], test_cg_niter[-1]),
    (dict_symbols_CG["shift"], test_cg_shift[-1]),
]

/root/src/lib/lab_lib.py:8785: UserWarning: ./csv_files/cg_na75000_nonzer18_niter30_shift60_process4.csv is empty.
  warnings.warn(f"{filePath} is empty.")
/root/src/lib/lab_lib.py:8785: UserWarning: ./csv_files/cg_na75000_nonzer15_niter90_shift80_process8.csv is empty.
  warnings.warn(f"{filePath} is empty.")


_list_series: list[pd.Series] = []

for functionNames in functionNames_CG:
    pass


def ret_averaged_rawDF_cg(
    list_process=list[int],
    list_na=list[int],
    list_nonzer=list[int],
    list_niter=list[int],
    list_shift=list[int],
    list_csvDir=list[str],
    resVar=str,
):
    """複数のCSVからDFを取得する関数

    引数resVarで指定された列がInclusiveもしくはExclusiveの場合はそれらが秒に変換され、InclusivePerCall列もしくはExclusivePerCall列が生成される

    Args:
        list_process(list[int]):プロセス数
        list_na(list[int]):初期変数na
        list_nonzer(list[int]):初期変数nonzer
        list_niter(list[int]):初期変数niter
        list_shift(list[int]):初期変数shift
        list_csvDir(list[str]):CSVファイルのディレクトリを複数保持したリスト
        resVar(str):目的変数のカラム名
    """
    pass


%reset -f

# jupyter_pwd = %pwd
# if jupyter_pwd == "/":
#     %cd /workspace


# ipynb形式のライブラリのインポート
%run ./lib/lib.ipynb

# 生データの入ったCSVファイルの保持されたディレクトリ名を格納している変数
csvDirPath = "./csv_files/"

# NPBのベンチマーク名のリスト
benchmarkNames = ["cg", "ep", "ft", "is", "lu", "mg"]

# LULESH ベンチマークプログラムのプロセス数・問題サイズ・イテレーション数
lulesh_processes: list[int] = [8, 27, 64, 125, 216, 343, 512]
lulesh_iterations: list[int] = [8, 16, 32, 64, 128, 256]
lulesh_sizes: list[int] = [16, 24, 32, 48, 64, 128]

# ipynb形式のライブラリノートを.py形式に変更したものをインポート
import lib
import lib.lab_lib
from lib.lab_lib import *

train_lulesh_processes: list[int] = [8, 27, 64, 125, 216, 343]
train_lulesh_iterations: list[int] = [8, 16, 32, 64, 128]
train_lulesh_sizes: list[int] = [16, 24, 32, 48]

test_lulesh_processes: list[int] = [512, 729, 1000]
test_lulesh_iterations: list[int] = [256, 512, 1024]
test_lulesh_sizes: list[int] = [64, 96, 128]

list_modelName: list[str] = [
    "modelIp",
    "modelLog",
    "modelLinAndIp",
    "modelLinAndLog",
    "modelIpAndLin",
    "modelIpAndLog",
    "modelLogAndLin",
    "modelLogAndIp",
    "modelProcessDividedByProblemSize",
    "modelProblemSizeDividedByProcess",
    "modelInfiniteProductOfProblemSizeMultipliedByProcesses",
    "modelInfiniteProductOfProblemSizeDividedByProcesses",
    "modelLinearSumOf2elementCombination",
    "modelLinearSumOfElementCombinations",
    "modelLinearSumOf2elementCombinationWithSquared",
    "modelLinearSumOf2elementCombinationWithCubed",
    "modelSquareRootOfProcess",
    "modelSquareRootTimesOtherElems",
    "modelObeyOneParameter",
    "modelLin"
    # "modelBasicTree",
]


# ipynb形式のライブラリのインポート
%run ./lib/lib.ipynb

# ipynb形式のライブラリノートを.py形式に変更したものをインポート
import lib
import lib.lab_lib
from lib.lab_lib import *

# 生データの入ったCSVファイルの保持されたディレクトリ名を格納している変数
csvDirPath = "./csv_files/"

# NPBのベンチマーク名のリスト
benchmarkNames = ["cg", "ep", "ft", "is", "lu", "mg"]

# NPBのプロセス数
npb_process: list[int] = [2, 4, 8, 16, 32, 64, 128, 256]
train_npb_process: list[int] = npb_process[:-1]
test_npb_process: list[int] = npb_process[-1:]
# NPBのCGの初期変数
cg_na: list[int] = [14000, 30000, 75000, 100000, 1500000]
cg_nonzer: list[int] = [11, 12, 13, 14, 15, 18, 21]
cg_niter: list[int] = [15, 30, 75, 90, 100]
cg_shift: list[int] = [20, 40, 60, 80, 110, 200]

train_cg_na: list[int] = cg_na[:-1]
train_cg_nonzer: list[int] = cg_nonzer[:-1]
train_cg_niter: list[int] = cg_niter[:-1]
train_cg_shift: list[int] = cg_shift[:-1]

test_cg_na: list[int] = cg_na[-1:]
test_cg_nonzer: list[int] = cg_nonzer[-1:]
test_cg_niter: list[int] = cg_niter[-1:]
test_cg_shift: list[int] = cg_shift[-1:]

# LULESH ベンチマークプログラムのプロセス数・問題サイズ・イテレーション数
lulesh_processes: list[int] = [8, 27, 64, 125, 216, 343, 512]
lulesh_iterations: list[int] = [8, 16, 32, 64, 128, 256]
lulesh_sizes: list[int] = [16, 24, 32, 48, 64, 128]

train_lulesh_processes: list[int] = [8, 27, 64, 125, 216, 343]
train_lulesh_iterations: list[int] = [8, 16, 32, 64, 128]
train_lulesh_sizes: list[int] = [16, 24, 32, 48]

test_lulesh_processes: list[int] = [512, 729, 1000]
test_lulesh_iterations: list[int] = [256, 512, 1024]
test_lulesh_sizes: list[int] = [64, 96, 128]

# Extra-Pのオプション
modelerNames: list[str] = [
    # "refining",
    "multi-parameter",
    "default",
    # "basic --options poly_exponents=-1,0,1,2,3 log_exponents=0,1 force_combination_exponents=1 allow_negative_exponents=1"
]

modelerOption: str = """ --options \#spm=Basic \#spo=poly_exponents=-1,0,1,2,3,log_exponents=0,1,force_combination_exponents=1,allow_negative_exponents=True"""

list_modelName: list[str] = [
    "modelIp",
    "modelLog",
    "modelLinAndIp",
    "modelLinAndLog",
    "modelIpAndLin",
    "modelIpAndLog",
    "modelLogAndLin",
    "modelLogAndIp",
    "modelProcessDividedByProblemSize",
    "modelProblemSizeDividedByProcess",
    "modelInfiniteProductOfProblemSizeMultipliedByProcesses",
    "modelInfiniteProductOfProblemSizeDividedByProcesses",
    "modelLinearSumOf2elementCombination",
    "modelLinearSumOfElementCombinations",
    "modelLinearSumOf2elementCombinationWithSquared",
    "modelLinearSumOf2elementCombinationWithCubed",
    "modelSquareRootOfProcess",
    "modelSquareRootTimesOtherElems",
    "modelObeyOneParameter",
    "modelLin"
    # "modelBasicTree",
]
list_csvDir = [
    "./csv_files/lulesh_1st/",
    "./csv_files/lulesh_2nd/",
    "./csv_files/lulesh_3rd/",
]

DEBUG:__main__:hello
DEBUG:__main__:hello


# LULESH


expVar: list[str] = ["process", "iteration", "size"]
resVar: str = "Exclusive"

trainDF: pd.DataFrame = ret_averaged_rawDF_lulesh(
    list_process=train_lulesh_processes,
    list_iteration=train_lulesh_iterations,
    list_size=train_lulesh_sizes,
    list_csvDir=list_csvDir,
    resVar=resVar,
)

testDF: pd.DataFrame = ret_averaged_rawDF_lulesh(
    list_process=test_lulesh_processes,
    list_iteration=test_lulesh_iterations,
    list_size=test_lulesh_sizes,
    list_csvDir=list_csvDir,
    resVar=resVar,
)

testDF = testDF.reset_index()
trainDF = trainDF.reset_index()

testDF["functionName"] = testDF["Name"]
trainDF["functionName"] = trainDF["Name"]

functionNames: list[str] = sorted(list(set(trainDF["Name"])))

dict_symbols = {}
for elem in expVar:
    dict_symbols[elem] = symbols(elem, real=True)

target_env = [
    (dict_symbols["size"], test_lulesh_sizes[-1]),
    (dict_symbols["iteration"], test_lulesh_iterations[-1]),
    (dict_symbols["process"], test_lulesh_processes[-1]),
]


# Extra-P による関数コール回数予測

resVar: str = "#Call"
expVar: list[str] = ["process", "iteration", "size"]

dict_models_perFunc: dict[str, dict[str, str]] = {}

for functionName in functionNames:

    trainDF_perFunc: pd.DataFrame = trainDF[
        trainDF["Name"] == functionName
    ].reset_index()

    str_ExtraPinputData_call: str = gen_ExtraPinputDataFromDF(
        inputDF=trainDF_perFunc,
        expVar=expVar,
        resVar=resVar,
    )

    # モデル格納用の辞書
    dictModels: dict[str, str] = {}

    # Extra-P によるモデルの構築
    converted_functionName: str = re.sub(
        r'[\\|/|:|?|.|"|<|>|\|\(|\)|]', "-", functionName
    )
    fileName: str = f"input_lulesh_@{converted_functionName}"
    fileDir: str = f"./extra-p_docker/share/"
    filePath: str = fileDir + fileName
    print(f"functionName = {functionName}")
    print(f"converted_functionName = {converted_functionName}")
    print(f"filePath = {filePath}")
    with open(file=filePath, mode="w") as f:
        f.write(str_ExtraPinputData_call)

    for modelerName in modelerNames:
        print(f"extrap --text {filePath} --modeler {modelerName} {modelerOption}")
        str_resFromExtraP: str = subprocess.run(
            f"extrap --text {filePath} --modeler {modelerName} {modelerOption}  | grep Model",
            stdout=subprocess.PIPE,
            text=True,
            shell=True,
        ).stdout
        str_resFromExtraP = str_resFromExtraP.replace("Model: ", "")
        str_resFromExtraP = convert_log(str_resFromExtraP)
        model_fromExtraP = sympify(str_resFromExtraP, locals=dict_symbols)

        dictModels[modelerName] = model_fromExtraP

    # 提案手法によるモデルの構築
    bestModelDict: dict = return_bestModelObject(
        inputDF=trainDF_perFunc,
        list_expVar=expVar,
        list_resVar=[resVar],
        list_modelName=list_modelName,
    )

    dictModels["proposal"] = bestModelDict["object"]

    dict_models_perFunc[functionName] = dictModels

functionName = .TAU_application
converted_functionName = -TAU_application
filePath = ./extra-p_docker/share/input_lulesh_@-TAU_application
extrap --text ./extra-p_docker/share/input_lulesh_@-TAU_application --modeler multi-parameter  --options \#spm=Basic \#spo=poly_exponents=-1,0,1,2,3,log_exponents=0,1,force_combination_exponents=1,allow_negative_exponents=True

Loading file: 100%|██████████| [00:00<00:00, Validating experiment]
Generating models: |          | [00:00<?]/usr/local/lib/python3.10/site-packages/extrap/modelers/single_parameter/basic.py:273: UserWarning: Number of measurements for a parameter needs to be at least 5 in order to create a performance model.
  warnings.warn(
Generating models: 100%|██████████| [00:00<00:00]

extrap --text ./extra-p_docker/share/input_lulesh_@-TAU_application --modeler default  --options \#spm=Basic \#spo=poly_exponents=-1,0,1,2,3,log_exponents=0,1,force_combination_exponents=1,allow_negative_exponents=True

Loading file: 100%|██████████| [00:00<00:00, Validating experiment]
Generating models: |          | [00:00<?]/usr/local/lib/python3.10/site-packages/extrap/modelers/single_parameter/basic.py:273: UserWarning: Number of measurements for a parameter needs to be at least 5 in order to create a performance model.
  warnings.warn(
Generating models: 100%|██████████| [00:00<00:00]
/usr/local/lib/python3.10/site-packages/scipy/optimize/_minpack_py.py:881: OptimizeWarning: Covariance of the parameters could not be estimated
  warnings.warn('Covariance of the parameters could not be estimated',
/usr/local/lib/python3.10/site-packages/scipy/optimize/_minpack_py.py:881: OptimizeWarning: Covariance of the parameters could not be estimated
  warnings.warn('Covariance of the parameters could not be estimated',
/usr/local/lib/python3.10/site-packages/scipy/optimize/_minpack_py.py:881: OptimizeWarning: Covariance of the parameters could not be estimated
  warnings.warn('Covariance of the parameters could not be estimated',

functionName = MPI_Allreduce()
converted_functionName = MPI_Allreduce--
filePath = ./extra-p_docker/share/input_lulesh_@MPI_Allreduce--
extrap --text ./extra-p_docker/share/input_lulesh_@MPI_Allreduce-- --modeler multi-parameter  --options \#spm=Basic \#spo=poly_exponents=-1,0,1,2,3,log_exponents=0,1,force_combination_exponents=1,allow_negative_exponents=True

Loading file: 100%|██████████| [00:00<00:00, Validating experiment]
Generating models: |          | [00:00<?]/usr/local/lib/python3.10/site-packages/extrap/modelers/single_parameter/basic.py:273: UserWarning: Number of measurements for a parameter needs to be at least 5 in order to create a performance model.
  warnings.warn(
Generating models: 100%|██████████| [00:00<00:00]

extrap --text ./extra-p_docker/share/input_lulesh_@MPI_Allreduce-- --modeler default  --options \#spm=Basic \#spo=poly_exponents=-1,0,1,2,3,log_exponents=0,1,force_combination_exponents=1,allow_negative_exponents=True

Loading file: 100%|██████████| [00:00<00:00, Validating experiment]
Generating models: |          | [00:00<?]/usr/local/lib/python3.10/site-packages/extrap/modelers/single_parameter/basic.py:273: UserWarning: Number of measurements for a parameter needs to be at least 5 in order to create a performance model.
  warnings.warn(
Generating models: 100%|██████████| [00:00<00:00]

functionName = MPI_Barrier()
converted_functionName = MPI_Barrier--
filePath = ./extra-p_docker/share/input_lulesh_@MPI_Barrier--
extrap --text ./extra-p_docker/share/input_lulesh_@MPI_Barrier-- --modeler multi-parameter  --options \#spm=Basic \#spo=poly_exponents=-1,0,1,2,3,log_exponents=0,1,force_combination_exponents=1,allow_negative_exponents=True

Loading file: 100%|██████████| [00:00<00:00, Validating experiment]
Generating models: |          | [00:00<?]/usr/local/lib/python3.10/site-packages/extrap/modelers/single_parameter/basic.py:273: UserWarning: Number of measurements for a parameter needs to be at least 5 in order to create a performance model.
  warnings.warn(
Generating models: 100%|██████████| [00:00<00:00]


_list_series: list[pd.Series] = []

for functionName in functionNames:
    pass
    testDF_perFunc: pd.DataFrame = testDF[testDF["Name"] == functionName].reset_index()

    tmp_series: pd.Series = pd.Series(dtype="object")
    # Extra-P による予測
    for modelerName in modelerNames:
        tmp_series[modelerName] = (
            dict_models_perFunc[functionName][modelerName].subs(target_env).evalf()
        )
    # 提案手法による予測
    _tmp_testDF_perFunc: pd.DataFrame = testDF_perFunc[
        (testDF_perFunc["process"] == test_lulesh_processes[-1])
        & (testDF_perFunc["iteration"] == test_lulesh_iterations[-1])
        & (testDF_perFunc["size"] == test_lulesh_sizes[-1])
        & (testDF_perFunc["Name"] == functionName)
    ]
    try:
        tmp_series["proposal"] = float(
            np.array(
                dict_models_perFunc[functionName]["proposal"].predict(
                    inputDF=_tmp_testDF_perFunc[expVar]
                )
            )
        )
    except:
        warnings.warn(f"{functionName} has no data at testDF")
        continue
    # 実測値
    tmp_series["real_call"] = _tmp_testDF_perFunc.reset_index(drop=True).loc[0]["#Call"]
    tmp_series["functionName"] = functionName

    _list_series.append(tmp_series)


DF_about_call_at_lulesh: pd.DataFrame = pd.DataFrame(_list_series)
DF_about_call_at_lulesh = add_relativeErrorRateCol(
    inputDF=DF_about_call_at_lulesh,
    real_colName="real_call",
    predicted_colName="default",
    targetColName="real_call vs default",
)
# DF_about_call_at_lulesh = add_relativeErrorRateCol(
#     inputDF=DF_about_call_at_lulesh,
#     real_colName="real_call",
#     predicted_colName="basic",
#     targetColName="real_call vs basic",
# )
# DF_about_call_at_lulesh = add_relativeErrorRateCol(
#     inputDF=DF_about_call_at_lulesh,
#     real_colName="real_call",
#     predicted_colName="refining",
#     targetColName="real_call vs refining",
# )
DF_about_call_at_lulesh = add_relativeErrorRateCol(
    inputDF=DF_about_call_at_lulesh,
    real_colName="real_call",
    predicted_colName="multi-parameter",
    targetColName="real_call vs multi-parameter",
)
DF_about_call_at_lulesh = add_relativeErrorRateCol(
    inputDF=DF_about_call_at_lulesh,
    real_colName="real_call",
    predicted_colName="proposal",
    targetColName="real_call vs proposal",
)


print(DF_about_call_at_lulesh.to_csv())

,multi-parameter,default,proposal,real_call,functionName,real_call vs default,real_call vs multi-parameter,real_call vs proposal
0,0.999999999999999,0.999999999999999,1.0,1.0,.TAU_application,1.11022302462516e-13,1.11022302462516e-13,0.0
1,1023.00000000000,1023.00000000000,1023.0,1023.0,MPI_Allreduce(),2.22261657324763e-14,2.22261657324763e-14,0.0
2,0.999999999999999,0.999999999999999,1.0,1.0,MPI_Barrier(),1.11022302462516e-13,1.11022302462516e-13,0.0
3,9221.00000000000,9221.00000000000,9221.000000000004,9221.0,MPI_Comm_rank(),1.97265958523572e-14,1.97265958523572e-14,3.945319170471438e-14
4,0.999999999999999,0.999999999999999,1.0,1.0,MPI_Comm_size(),1.11022302462516e-13,1.11022302462516e-13,0.0
5,0.999999999999999,0.999999999999999,1.0,1.0,MPI_Finalize(),1.11022302462516e-13,1.11022302462516e-13,0.0
6,0.999999999999999,0.999999999999999,1.0,1.0,MPI_Init(),1.11022302462516e-13,1.11022302462516e-13,0.0
7,26764.7475504032,26764.7475504032,51757.94232139088,37732.8,MPI_Irecv(),29.0676876605944,29.0676876605944,37.16963045782682
8,26764.7475504032,26764.7475504032,51757.94232139088,37732.8,MPI_Isend(),29.0676876605944,29.0676876605944,37.16963045782682
9,0.999999999999999,0.999999999999999,1.0,1.0,MPI_Reduce(),1.11022302462516e-13,1.11022302462516e-13,0.0
10,26764.7475504032,26764.7475504032,51757.94232139088,37732.8,MPI_Wait(),29.0676876605944,29.0676876605944,37.16963045782682
11,3073.00000000000,3073.00000000000,3073.0000000000005,3073.0,MPI_Waitall(),4.43944696602471e-14,4.43944696602471e-14,1.479815655341569e-14
12,2149585472.22544,2149585472.22544,2073910598.7216053,2149580000.0,Real_t_CalcElemVolume(const,0.000254571844002869,0.000254571844002869,3.520194702146219
13,2.00000000000000,2.00000000000000,2.0,2.0,StrToInt,1.11022302462516e-13,1.11022302462516e-13,0.0
14,0.999999999999999,0.999999999999999,1.0,1.0,int_main(int_char_**),1.11022302462516e-13,1.11022302462516e-13,0.0
15,1024.00000000000,1024.00000000000,1024.0,1024.0,void_CalcKinematicsForElems(Domain,0,0,0.0
16,1024.00000000000,1024.00000000000,1024.0,1024.0,void_CommMonoQ(Domain,0,0,0.0
17,3073.00000000000,3073.00000000000,3073.0000000000005,3073.0,void_CommRecv(Domain,4.43944696602471e-14,4.43944696602471e-14,1.479815655341569e-14
18,1025.00000000000,1025.00000000000,1025.0,1025.0,void_CommSBN(Domain,2.21827976042178e-14,2.21827976042178e-14,0.0
19,3073.00000000000,3073.00000000000,3073.0000000000005,3073.0,void_CommSend(Domain,4.43944696602471e-14,4.43944696602471e-14,1.479815655341569e-14
20,1024.00000000000,1024.00000000000,1024.0,1024.0,void_CommSyncPosVel(Domain,0,0,0.0
21,0.999999999999999,0.999999999999999,1.0,1.0,void_Domain::BuildMesh(Int_t_Int_t_Int_t),1.11022302462516e-13,1.11022302462516e-13,0.0
22,0.999999999999999,0.999999999999999,1.0,1.0,void_Domain::CreateRegionIndexSets(Int_t_Int_t),1.11022302462516e-13,1.11022302462516e-13,0.0
23,0.999999999999999,0.999999999999999,1.0,1.0,void_Domain::Domain(Int_t_Index_t_Index_t_Index_t_Index_t_Int_t_Int_t_Int_t_Int_t),1.11022302462516e-13,1.11022302462516e-13,0.0
24,0.999999999999999,0.999999999999999,1.0,1.0,void_Domain::SetupBoundaryConditions(Int_t),1.11022302462516e-13,1.11022302462516e-13,0.0
25,0.999999999999999,0.999999999999999,1.0,1.0,void_Domain::SetupCommBuffers(Int_t),1.11022302462516e-13,1.11022302462516e-13,0.0
26,0.999999999999999,0.999999999999999,1.0,1.0,void_Domain::SetupElementConnectivities(Int_t),1.11022302462516e-13,1.11022302462516e-13,0.0
27,0.999999999999999,0.999999999999999,1.0,1.0,void_Domain::SetupSymmetryPlanes(Int_t),1.11022302462516e-13,1.11022302462516e-13,0.0
28,0.999999999999999,0.999999999999999,1.0,1.0,void_Domain::~Domain(),1.11022302462516e-13,1.11022302462516e-13,0.0
29,0.999999999999999,0.999999999999999,1.0,1.0,void_InitMeshDecomp(Int_t_Int_t_Int_t,1.11022302462516e-13,1.11022302462516e-13,0.0
30,0.999999999999999,0.999999999999999,1.0,1.0,void_ParseCommandLineOptions(int_char_**_Int_t_cmdLineOpts,1.11022302462516e-13,1.11022302462516e-13,0.0
31,0.000999993948753132,0.000999993948753132,0.0009999939487532368,0.001,void_VerifyAndWriteFinalOutput(Real_t_Domain,0.000605124686817171,0.000605124686817171,0.0006051246763220941


DF_about_call_at_lulesh.mean()

/tmp/ipykernel_121/1662607098.py:1: FutureWarning: The default value of numeric_only in DataFrame.mean is deprecated. In a future version, it will default to False. In addition, specifying 'numeric_only=None' is deprecated. Select only valid columns or specify the value of numeric_only to silence this warning.
  DF_about_call_at_lulesh.mean()

multi-parameter                 6.717779e+07
default                         6.717779e+07
proposal                        6.481530e+07
real_call                       6.717865e+07
real_call vs default            2.725123e+00
real_call vs multi-parameter    2.725123e+00
real_call vs proposal           3.594678e+00
dtype: float64


expVar: list[str] = ["process", "iteration", "size"]
# resVar: str = "Exclusive"

# trainDF_lulesh: pd.DataFrame = ret_averaged_rawDF_lulesh(
#     list_process=train_lulesh_processes,
#     list_iteration=train_lulesh_iterations,
#     list_size=train_lulesh_sizes,
#     list_csvDir=list_csvDir,
#     resVar=resVar,
# )

# testDF_lulesh: pd.DataFrame = ret_averaged_rawDF_lulesh(
#     list_process=test_lulesh_processes,
#     list_iteration=test_lulesh_iterations,
#     list_size=test_lulesh_sizes,
#     list_csvDir=list_csvDir,
#     resVar=resVar,
# )

# testDF_lulesh = testDF_lulesh.reset_index()
# trainDF_lulesh = trainDF_lulesh.reset_index()

# testDF_lulesh["functionName"] = testDF_lulesh["Name"]
# trainDF_lulesh["functionName"] = trainDF_lulesh["Name"]

# functionNames_lulesh: list[str] = sorted(list(set(trainDF["Name"])))

dict_symbols_lulesh = {}
for elem in expVar:
    dict_symbols_lulesh[elem] = symbols(elem, real=True)

target_env_lulesh = [
    (dict_symbols_lulesh["size"], test_lulesh_sizes[-1]),
    (dict_symbols_lulesh["iteration"], test_lulesh_iterations[-1]),
    (dict_symbols_lulesh["process"], test_lulesh_processes[-1]),
]

modelerName: str = "multi-parameter"
modelerOption: str = """ --options \#spm=Basic \#spo=poly_exponents=-1,0,1,2,3,log_exponents=0,1,force_combination_exponents=1,allow_negative_exponents=True"""


%reset -f

# jupyter_pwd = %pwd
# if jupyter_pwd == "/":
#     %cd /workspace


# ipynb形式のライブラリのインポート
%run ./lib/lib.ipynb

# 生データの入ったCSVファイルの保持されたディレクトリ名を格納している変数
csvDirPath = "./csv_files/"

# NPBのベンチマーク名のリスト
benchmarkNames = ["cg", "ep", "ft", "is", "lu", "mg"]

# LULESH ベンチマークプログラムのプロセス数・問題サイズ・イテレーション数
lulesh_processes: list[int] = [8, 27, 64, 125, 216, 343, 512]
lulesh_iterations: list[int] = [8, 16, 32, 64, 128, 256]
lulesh_sizes: list[int] = [16, 24, 32, 48, 64, 128]

# ipynb形式のライブラリノートを.py形式に変更したものをインポート
import lib
import lib.lab_lib
from lib.lab_lib import *

train_lulesh_processes: list[int] = [8, 27, 64, 125, 216, 343]
train_lulesh_iterations: list[int] = [8, 16, 32, 64, 128]
train_lulesh_sizes: list[int] = [16, 24, 32, 48]

test_lulesh_processes: list[int] = [512, 729, 1000]
test_lulesh_iterations: list[int] = [256, 512, 1024]
test_lulesh_sizes: list[int] = [64, 96, 128]

list_modelName: list[str] = [
    "modelIp",
    "modelLog",
    "modelLinAndIp",
    "modelLinAndLog",
    "modelIpAndLin",
    "modelIpAndLog",
    "modelLogAndLin",
    "modelLogAndIp",
    "modelProcessDividedByProblemSize",
    "modelProblemSizeDividedByProcess",
    "modelInfiniteProductOfProblemSizeMultipliedByProcesses",
    "modelInfiniteProductOfProblemSizeDividedByProcesses",
    "modelLinearSumOf2elementCombination",
    "modelLinearSumOfElementCombinations",
    "modelLinearSumOf2elementCombinationWithSquared",
    "modelLinearSumOf2elementCombinationWithCubed",
    "modelSquareRootOfProcess",
    "modelSquareRootTimesOtherElems",
    "modelObeyOneParameter",
    "modelLin"
    # "modelBasicTree",
]


# ipynb形式のライブラリのインポート
%run ./lib/lib.ipynb

# ipynb形式のライブラリノートを.py形式に変更したものをインポート
import lib
import lib.lab_lib
from lib.lab_lib import *

# 生データの入ったCSVファイルの保持されたディレクトリ名を格納している変数
csvDirPath = "./csv_files/"

# NPBのベンチマーク名のリスト
benchmarkNames = ["cg", "ep", "ft", "is", "lu", "mg"]

# NPBのプロセス数
npb_process: list[int] = [2, 4, 8, 16, 32, 64, 128, 256]
train_npb_process: list[int] = npb_process[:-1]
test_npb_process: list[int] = npb_process[-1:]
# NPBのCGの初期変数
cg_na: list[int] = [14000, 30000, 75000, 100000, 1500000]
cg_nonzer: list[int] = [11, 12, 13, 14, 15, 18, 21]
cg_niter: list[int] = [15, 30, 75, 90, 100]
cg_shift: list[int] = [20, 40, 60, 80, 110, 200]

train_cg_na: list[int] = cg_na[:-1]
train_cg_nonzer: list[int] = cg_nonzer[:-1]
train_cg_niter: list[int] = cg_niter[:-1]
train_cg_shift: list[int] = cg_shift[:-1]

test_cg_na: list[int] = cg_na[-1:]
test_cg_nonzer: list[int] = cg_nonzer[-1:]
test_cg_niter: list[int] = cg_niter[-1:]
test_cg_shift: list[int] = cg_shift[-1:]

# LULESH ベンチマークプログラムのプロセス数・問題サイズ・イテレーション数
lulesh_processes: list[int] = [8, 27, 64, 125, 216, 343, 512]
lulesh_iterations: list[int] = [8, 16, 32, 64, 128, 256]
lulesh_sizes: list[int] = [16, 24, 32, 48, 64, 128]

train_lulesh_processes: list[int] = [8, 27, 64, 125, 216, 343]
train_lulesh_iterations: list[int] = [8, 16, 32, 64, 128]
train_lulesh_sizes: list[int] = [16, 24, 32, 48]

test_lulesh_processes: list[int] = [512, 729, 1000]
test_lulesh_iterations: list[int] = [256, 512, 1024]
test_lulesh_sizes: list[int] = [64, 96, 128]

# Extra-Pのオプション
modelerNames: list[str] = [
    # "refining",
    "multi-parameter",
    "default",
    # "basic --options poly_exponents=-1,0,1,2,3 log_exponents=0,1 force_combination_exponents=1 allow_negative_exponents=1"
]

modelerOption: str = """ --options \#spm=Basic \#spo=poly_exponents=-1,0,1,2,3,log_exponents=0,1,force_combination_exponents=1,allow_negative_exponents=True"""

list_modelName: list[str] = [
    "modelIp",
    "modelLog",
    "modelLinAndIp",
    "modelLinAndLog",
    "modelIpAndLin",
    "modelIpAndLog",
    "modelLogAndLin",
    "modelLogAndIp",
    "modelProcessDividedByProblemSize",
    "modelProblemSizeDividedByProcess",
    "modelInfiniteProductOfProblemSizeMultipliedByProcesses",
    "modelInfiniteProductOfProblemSizeDividedByProcesses",
    "modelLinearSumOf2elementCombination",
    "modelLinearSumOfElementCombinations",
    "modelLinearSumOf2elementCombinationWithSquared",
    "modelLinearSumOf2elementCombinationWithCubed",
    "modelSquareRootOfProcess",
    "modelSquareRootTimesOtherElems",
    "modelObeyOneParameter",
    "modelLin"
    # "modelBasicTree",
]
list_csvDir = [
    "./csv_files/lulesh_1st/",
    "./csv_files/lulesh_2nd/",
    "./csv_files/lulesh_3rd/",
]

DEBUG:__main__:hello
DEBUG:__main__:hello


# コストのグラフの作成

_list_series: list[pd.Series] = []
pis: list[str]

for elem_process in train_lulesh_processes + test_lulesh_processes:
    fileDir: str = "./txt_files/ElapFiles/"
    fileName: str = f"p{elem_process}_Elap"
    with open(fileDir + fileName) as f:
        l: list[str] = [s.strip() for s in f.readlines()]
    if len(l) % 2 != 0:
        warnings.warn("ファイルの行数が偶数ではありません")
    for i in range(len(l)):
        if i % 2 == 0:
            pis = l[i].split(sep="/")[-3:]
        else:
            time = l[i].split(sep=" ")[-2]
            _series: pd.Series = pd.Series(
                {
                    "p": int(pis[0].replace("p", "")),
                    "i": int(pis[1].replace("i", "")),
                    "s": int(pis[2].replace("s", "")),
                    "time": float(time),
                }
            )
            _series["cost"] = _series["p"] * _series["time"]
            _list_series.append(_series)
        # print(f"pis={pis}, time={time}")

DF_pis_time: pd.DataFrame = pd.DataFrame(data=_list_series)

target_p_border: int = test_lulesh_processes[0]
target_i_border: int = test_lulesh_iterations[0]
target_s_border: int = test_lulesh_sizes[0]

DF_pis_time_at_train: pd.DataFrame = DF_pis_time[
    (DF_pis_time["p"] < target_p_border)
    & (DF_pis_time["i"] < target_i_border)
    & (DF_pis_time["s"] < target_s_border)
]
DF_pis_time_at_target: pd.DataFrame = DF_pis_time[
    (DF_pis_time["p"] >= target_p_border)
    & (DF_pis_time["i"] >= target_i_border)
    & (DF_pis_time["s"] >= target_s_border)
]

cost_to_build_model: float = sum(DF_pis_time_at_train["cost"])

DF_pis_time_at_target = DF_pis_time_at_target.sort_values(by=["p", "i", "s"])

DF_pis_time_at_train = DF_pis_time_at_train.sort_values(by=["p", "i", "s"])

condition_col: list[str] = ["p", "i", "s"]
averaged_DF_pis_time_at_target = DF_pis_time_at_target.groupby(condition_col).aggregate(
    np.mean
)

trace_data = []
layout = go.Layout(barmode="stack")
for i, sr in DF_pis_time_at_target.reset_index().iterrows():
    x = [j for j in range(1, len(DF_pis_time_at_target) + 1)]
    y = [0] * len(DF_pis_time_at_target)
    for k in range(int(i), len(DF_pis_time_at_target)):
        y[k] = sr["cost"]
    trace = go.Bar(x=x, y=y, name=f"p{int(sr['p'])}i{int(sr['i'])}s{int(sr['s'])}")
    trace_data.append(trace)

fig = go.Figure(data=trace_data, layout=layout)
fig.add_shape(
    type="line",
    x0=0,
    x1=len(DF_pis_time_at_target),
    y0=cost_to_build_model,
    y1=cost_to_build_model,
    line=dict(
        color="Red",
    ),
    xref="x",
    yref="y",
    name="モデル構築に要するコスト",
)
fig.update_layout(xaxis=dict(title="予測対象の個数"), yaxis=dict(title="コスト"))
fig.show()


averaged_DF_pis_time_at_target = averaged_DF_pis_time_at_target.reset_index()

graph_x: list[int] = list(range(1, len(averaged_DF_pis_time_at_target) + 1))
graph_y: list[float] = [0] * len(averaged_DF_pis_time_at_target)

for i, sr in averaged_DF_pis_time_at_target.iterrows():
    for k in range(i, len(averaged_DF_pis_time_at_target)):
        graph_y[k] += sr["cost"]

graph_x_to_build: list[float] = list(range(len(averaged_DF_pis_time_at_target)))
graph_x_to_build[-1] = 27
graph_y_to_build: list[float] = [cost_to_build_model] * len(
    averaged_DF_pis_time_at_target
)

symbols1: list[str] = ["実測"] * len(averaged_DF_pis_time_at_target)
symbols2: list[str] = ["予測"] * len(averaged_DF_pis_time_at_target)
_symbols: list[str] = symbols1 + symbols2

dataDF: pd.DataFrame = pd.DataFrame(
    data={
        "取得対象の実行規模の数": graph_x + graph_x_to_build,
        "コスト": graph_y + graph_y_to_build,
        "凡例": _symbols,
    }
)

fig = px.line(dataDF, x="取得対象の実行規模の数", y="コスト", color="凡例", log_y=True)
fig.update_layout(
    legend=dict(
        x=0.01,  # ①：X座標
        y=0.99,  # ①：Y座標
        xanchor="left",  # ②：X座標が凡例のどの部分を表すか
        yanchor="top",  # ②：Y座標が凡例のどの部分を表すか
        orientation="h",  # ③：凡例を横並びにする
    )
)
fig.show()

# data = go.Scatter(x=graph_x, y=graph_y, name = "予測対象のコスト", symbol="")
# fig = go.Figure(data=data)
# fig.add_shape(
#     type="line",
#     x0=0,
#     x1=len(averaged_DF_pis_time_at_target),
#     y0=cost_to_build_model,
#     y1=cost_to_build_model,
#     line=dict(
#         color="Red",
#     ),
#     xref="x",
#     yref="y",
#     name="モデル構築に必要なコスト",
# )
# fig.update_layout(
#     xaxis=dict(title="取得対象の実行規模の数"),
#     yaxis=dict(title="コスト", type="log"),
#     width=1000,
#     height=500,
# )
# fig.show()


print(dataDF.T.to_csv())

,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53
取得対象の実行規模の数,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,27
コスト,66560.0,307200.0,870400.0,1005226.6666666666,1493333.3333333333,2670933.333333333,2945706.6666666665,3918506.6666666665,6256640.0,6361130.0,6706190.0,7508090.0,7707350.0,8407190.0,10083890.0,10477550.0,11862650.0,15216050.0,15352716.666666666,15839383.333333332,17006050.0,17282716.666666668,18269383.333333336,20569383.333333336,21129383.333333336,23062716.666666668,27729383.333333336,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995,180397.09999999995
凡例,実測,実測,実測,実測,実測,実測,実測,実測,実測,実測,実測,実測,実測,実測,実測,実測,実測,実測,実測,実測,実測,実測,実測,実測,実測,実測,実測,予測,予測,予測,予測,予測,予測,予測,予測,予測,予測,予測,予測,予測,予測,予測,予測,予測,予測,予測,予測,予測,予測,予測,予測,予測,予測,予測


print(graph_y[-1])

print(cost_to_build_model)

print(100 * cost_to_build_model / graph_y[-1])

dataDF

27729383.333333336
180397.09999999995
0.6505629708077411


# 関数コール回数予測のためのモデル

resVar = "#Call"
expVar: list[str] = ["process", "iteration", "size"]


dict_symbols_lulesh = {}
for elem in expVar:
    dict_symbols_lulesh[elem] = symbols(elem, real=True)

target_env_lulesh = [
    (dict_symbols_lulesh["size"], test_lulesh_sizes[-1]),
    (dict_symbols_lulesh["iteration"], test_lulesh_iterations[-1]),
    (dict_symbols_lulesh["process"], test_lulesh_processes[-1]),
]

modelerName: str = "multi-parameter"
modelerOption: str = """ --options \#spm=Basic \#spo=poly_exponents=-1,0,1,2,3,log_exponents=0,1,force_combination_exponents=1,allow_negative_exponents=True"""

print(f"resVar = {resVar}")

trainDF_lulesh: pd.DataFrame = ret_averaged_rawDF_lulesh(
    list_process=train_lulesh_processes,
    list_iteration=train_lulesh_iterations,
    list_size=train_lulesh_sizes,
    list_csvDir=list_csvDir,
    resVar=resVar,
)

functionNames_lulesh: list[str] = sorted(list(set(trainDF_lulesh["Name"])))

dict_functionName_model_call: dict[str, str] = {}

for functionName in functionNames_lulesh:
    trainDF_perFunc: pd.DataFrame = trainDF_lulesh[
        trainDF_lulesh["Name"] == functionName
    ].reset_index()

    model_fromExtraP_all = get_ExtraP_model(
        inputDF_perFunc=trainDF_perFunc,
        expVar=expVar,
        resVar=resVar,
        functionName=functionName,
        benchmarkName="lulesh",
        modelerName=modelerName,
        modelerOption=modelerOption,
        dict_symbols=dict_symbols_lulesh,
    )
    dict_functionName_model_call[functionName] = model_fromExtraP_all

resVar = #Call


list_resVar: list[str] = ["Exclusive", "Inclusive", "#Subrs"]
dict_resVar_functionName_models_all: dict = {}
dict_resVar_functionName_models_perCall: dict = {}

for resVar in list_resVar:
    pass

    print(f"resVar = {resVar}")

    trainDF_lulesh: pd.DataFrame = ret_averaged_rawDF_lulesh(
        list_process=train_lulesh_processes,
        list_iteration=train_lulesh_iterations,
        list_size=train_lulesh_sizes,
        list_csvDir=list_csvDir,
        resVar=resVar,
    )

    print(f"{trainDF_lulesh.head()}")

    functionNames_lulesh: list[str] = sorted(list(set(trainDF_lulesh["Name"])))

    dict_functionName_model_all: dict[str, str] = {}
    dict_functionName_model_perCall: dict[str, str] = {}

    for functionName in functionNames_lulesh:
        trainDF_perFunc: pd.DataFrame = trainDF_lulesh[
            trainDF_lulesh["Name"] == functionName
        ].reset_index()

        model_fromExtraP_all = get_ExtraP_model(
            inputDF_perFunc=trainDF_perFunc,
            expVar=expVar,
            resVar=resVar,
            functionName=functionName,
            benchmarkName="lulesh",
            modelerName=modelerName,
            modelerOption=modelerOption,
            dict_symbols=dict_symbols_lulesh,
        )
        dict_functionName_model_all[functionName] = model_fromExtraP_all

        model_fromExtraP_perCall = get_ExtraP_model(
            inputDF_perFunc=trainDF_perFunc,
            expVar=expVar,
            resVar=f"{resVar}PerCall",
            functionName=functionName,
            benchmarkName="lulesh",
            modelerName=modelerName,
            modelerOption=modelerOption,
            dict_symbols=dict_symbols_lulesh,
        )
        dict_functionName_model_perCall[functionName] = model_fromExtraP_perCall

    dict_resVar_functionName_models_all[resVar] = dict_functionName_model_all
    dict_resVar_functionName_models_perCall[resVar] = dict_functionName_model_perCall

resVar = Exclusive
   %Time  Exclusive Inclusive  #Call     #Subrs  \
0  100.0   0.000003     4,647    1.0      1.000   
1  100.0   0.025000     4,647    1.0    100.125   
2   84.0   6.192000     3,901    1.0      0.000   
3   15.1   0.727667       699    1.0      0.000   
4    0.2   0.006000        11    8.0  32768.000   

                                 Name  process  iteration  size  \
0                    .TAU_application        8          8    16   
1               int_main(int_char_**)        8          8    16   
2                      MPI_Finalize()        8          8    16   
3                          MPI_Init()        8          8    16   
4  void_CalcKinematicsForElems(Domain        8          8    16   

   ExclusivePerCall  
0          0.000003  
1          0.025000  
2          3.901000  
3          0.699000  
4          0.000750  
resVar = Inclusive
   %Time Exclusive  Inclusive  #Call     #Subrs  \
0  100.0    0.0035   6.966333    1.0      1.000   
1  100.0        25   6.966333    1.0    100.125   
2   84.0     3,901   6.192000    1.0      0.000   
3   15.1       699   0.727667    1.0      0.000   
4    0.2         6   0.011000    8.0  32768.000   

                                 Name  process  iteration  size  \
0                    .TAU_application        8          8    16   
1               int_main(int_char_**)        8          8    16   
2                      MPI_Finalize()        8          8    16   
3                          MPI_Init()        8          8    16   
4  void_CalcKinematicsForElems(Domain        8          8    16   

   InclusivePerCall  
0          4.647000  
1          4.647000  
2          3.901000  
3          0.699000  
4          0.001375  
resVar = #Subrs
   %Time Exclusive Inclusive  #Call     #Subrs  \
0  100.0    0.0035     4,647    1.0      1.000   
1  100.0        25     4,647    1.0    100.125   
2   84.0     3,901     3,901    1.0      0.000   
3   15.1       699       699    1.0      0.000   
4    0.2         6        11    8.0  32768.000   

                                 Name  process  iteration  size  #SubrsPerCall  
0                    .TAU_application        8          8    16          1.000  
1               int_main(int_char_**)        8          8    16        100.125  
2                      MPI_Finalize()        8          8    16          0.000  
3                          MPI_Init()        8          8    16          0.000  
4  void_CalcKinematicsForElems(Domain        8          8    16       4096.000


dict_resultDF: dict[str, pd.DataFrame] = {}
for resVar in list_resVar:

    testDF_lulesh = ret_averaged_rawDF_lulesh(
        list_process=test_lulesh_processes,
        list_iteration=test_lulesh_iterations,
        list_size=test_lulesh_sizes,
        list_csvDir=list_csvDir,
        resVar=resVar,
    )

    print(f"{testDF_lulesh.head()}")

    functionNames_lulesh: list[str] = sorted(list(set(testDF_lulesh["Name"])))
    _list_series: list[pd.Series] = []

    for functionName in functionNames_lulesh:
        tmp_series: pd.Series = pd.Series(dtype="object")

        testDF_perFunc: pd.DataFrame = testDF_lulesh[
            (testDF_lulesh["Name"] == functionName)
            & (testDF_lulesh["process"] == test_lulesh_processes[-1])
            & (testDF_lulesh["iteration"] == test_lulesh_iterations[-1])
            & (testDF_lulesh["size"] == test_lulesh_sizes[-1])
        ].reset_index()

        if len(testDF_perFunc) == 1:
            tmp_series[f"real_{resVar}"] = testDF_perFunc.loc[0][resVar]
            tmp_series["functionName"] = functionName
            tmp_series[f"predicted_{resVar}"] = (
                dict_resVar_functionName_models_all[resVar][functionName]
                .subs(target_env_lulesh)
                .evalf()
            )

            # コール回数予測
            _predicted_call: float
            try:
                _predicted_call = (
                    dict_functionName_model_call[functionName]
                    .subs(target_env_lulesh)
                    .evalf()
                )
            except:
                _predicted_call = dict_functionName_model_call[functionName]

            # 予測対象値の１コール当たりの値
            _target_num_perCall: float = (
                dict_resVar_functionName_models_perCall[resVar][functionName]
                .subs(target_env_lulesh)
                .evalf()
            )

            # 予測されたコール回数 * 予測対象値の1コールあたりの値
            _predicted_num: float = _predicted_call * _target_num_perCall

            tmp_series[f"predicted_{resVar}_from_call"] = _predicted_num
            tmp_series["predicted_call"] = _predicted_call
            tmp_series["real_call"] = testDF_perFunc.loc[0][resVar]
            _list_series.append(tmp_series)

    print(resVar)
    _DF_tmp: pd.DataFrame = pd.DataFrame(data=_list_series)
    _DF_tmp = add_relativeErrorRateCol(
        inputDF=_DF_tmp,
        real_colName=f"real_{resVar}",
        predicted_colName=f"predicted_{resVar}",
        targetColName=f"real vs predicted all @{resVar}",
    )
    _DF_tmp = add_relativeErrorRateCol(
        inputDF=_DF_tmp,
        real_colName=f"real_{resVar}",
        predicted_colName=f"predicted_{resVar}_from_call",
        targetColName=f"real vs predicted using call @{resVar}",
    )
    dict_resultDF[resVar] = _DF_tmp

   %Time  Exclusive Inclusive  #Call      #Subrs  \
0  100.0   0.000004  2:13.310    1.0         1.0   
1  100.0  73.073333  2:13.310    1.0      2828.0   
2   17.9  14.006667    23,890  256.0  67108900.0   
3   14.6  18.807333    19,437  255.0         0.0   
4    7.7   3.580667    10,238  769.0     10503.8   

                                 Name  process  iteration  size  \
0                    .TAU_application      512        256    64   
1               int_main(int_char_**)      512        256    64   
2  void_CalcKinematicsForElems(Domain      512        256    64   
3                     MPI_Allreduce()      512        256    64   
4                void_CommSend(Domain      512        256    64   

   ExclusivePerCall  
0          0.000004  
1         72.675000  
2          0.054781  
3          0.076224  
4          0.000563  
Exclusive
   %Time Exclusive   Inclusive  #Call      #Subrs  \
0  100.0   0.00374  132.158333    1.0         1.0   
1  100.0  1:12.675  132.158333    1.0      2828.0   
2   17.9    14,024   23.865667  256.0  67108900.0   
3   14.6    19,437   18.807333  255.0         0.0   
4    7.7       433   10.272000  769.0     10503.8   

                                 Name  process  iteration  size  \
0                    .TAU_application      512        256    64   
1               int_main(int_char_**)      512        256    64   
2  void_CalcKinematicsForElems(Domain      512        256    64   
3                     MPI_Allreduce()      512        256    64   
4                void_CommSend(Domain      512        256    64   

   InclusivePerCall  
0        133.310000  
1        133.310000  
2          0.093320  
3          0.076224  
4          0.013313  
Inclusive
   %Time Exclusive Inclusive  #Call        #Subrs  \
0  100.0   0.00374  2:13.310    1.0  1.000000e+00   
1  100.0  1:12.675  2:13.310    1.0  2.828000e+03   
2   17.9    14,024    23,890  256.0  6.710890e+07   
3   14.6    19,437    19,437  255.0  0.000000e+00   
4    7.7       433    10,238  769.0  7.002533e+03   

                                 Name  process  iteration  size  #SubrsPerCall  
0                    .TAU_application      512        256    64       1.000000  
1               int_main(int_char_**)      512        256    64    2828.000000  
2  void_CalcKinematicsForElems(Domain      512        256    64  262144.140625  
3                     MPI_Allreduce()      512        256    64       0.000000  
4                void_CommSend(Domain      512        256    64      13.659038  
#Subrs


list_resVar: list[str] = ["Exclusive", "Inclusive", "#Subrs"]

print(dict_resultDF.keys())

print(dict_resultDF[list_resVar[0]].mean())

dict_resultDF[list_resVar[0]]

dict_keys(['Exclusive', 'Inclusive', '#Subrs'])
real_Exclusive                             1.453734e+02
predicted_Exclusive                        1.682026e+02
predicted_Exclusive_from_call              1.005981e+04
predicted_call                             6.352180e+07
real_call                                  1.453734e+02
real vs predicted all @Exclusive           9.842000e+01
real vs predicted using call @Exclusive    6.690583e+05
dtype: float64

/tmp/ipykernel_121/3090816874.py:5: FutureWarning:

The default value of numeric_only in DataFrame.mean is deprecated. In a future version, it will default to False. In addition, specifying 'numeric_only=None' is deprecated. Select only valid columns or specify the value of numeric_only to silence this warning.


print(dict_resultDF[list_resVar[1]].mean())
dict_resultDF[list_resVar[1]]

real_Inclusive                             3.516647e+02
predicted_Inclusive                        3.853701e+02
predicted_Inclusive_from_call              1.046825e+04
predicted_call                             6.352180e+07
real_call                                  3.516647e+02
real vs predicted all @Inclusive           1.170237e+02
real vs predicted using call @Inclusive    6.688036e+05
dtype: float64

/tmp/ipykernel_121/261304090.py:1: FutureWarning:

The default value of numeric_only in DataFrame.mean is deprecated. In a future version, it will default to False. In addition, specifying 'numeric_only=None' is deprecated. Select only valid columns or specify the value of numeric_only to silence this warning.


print(dict_resultDF[list_resVar[2]].mean())
dict_resultDF[list_resVar[2]]

real_#Subrs                             6.717856e+07+0.000000e+00j
predicted_#Subrs                        6.091424e+07+0.000000e+00j
predicted_#Subrs_from_call              1.192909e+08+0.000000e+00j
predicted_call                          6.352180e+07+0.000000e+00j
real_call                               6.717856e+07+0.000000e+00j
real vs predicted all @#Subrs          N000000000000a000000000000N
real vs predicted using call @#Subrs   N000000000000a000000000000N
dtype: complex128

/tmp/ipykernel_121/1563136227.py:1: FutureWarning:

The default value of numeric_only in DataFrame.mean is deprecated. In a future version, it will default to False. In addition, specifying 'numeric_only=None' is deprecated. Select only valid columns or specify the value of numeric_only to silence this warning.


jupyter_pwd = %pwd
if jupyter_pwd == "/":
    %cd /workspace


# ipynb形式のライブラリのインポート
%run ./lib/lib.ipynb

# ipynb形式のライブラリノートを.py形式に変更したものをインポート
import lib
import lib.lab_lib
from lib.lab_lib import *

# 生データの入ったCSVファイルの保持されたディレクトリ名を格納している変数
csvDirPath = "./csv_files/"

# NPBのベンチマーク名のリスト
benchmarkNames = ["cg", "ep", "ft", "is", "lu", "mg"]

# NPBのプロセス数
npb_process :list[int] = [2, 4, 8, 16, 32, 64, 128, 256]
train_npb_process :list[int] = npb_process[:-1]
test_npb_process :list[int] = npb_process[-1:]
# NPBのCGの初期変数
cg_na: list[int] = [14000, 30000, 75000, 100000, 1500000]
cg_nonzer: list[int] = [11, 12, 13, 14, 15, 18, 21]
cg_niter: list[int] = [15, 30, 75, 90, 100]
cg_shift: list[int] = [20, 40, 60, 80, 110, 200]

train_cg_na: list[int] = cg_na[:-1]
train_cg_nonzer: list[int] = cg_nonzer[:-1]
train_cg_niter: list[int] = cg_niter[:-1]
train_cg_shift: list[int] = cg_shift[:-1]

test_cg_na: list[int] = cg_na[-1:]
test_cg_nonzer: list[int] = cg_nonzer[-1:]
test_cg_niter: list[int] = cg_niter[-1:]
test_cg_shift: list[int] = cg_shift[-1:]

# LULESH ベンチマークプログラムのプロセス数・問題サイズ・イテレーション数
lulesh_processes: list[int] = [8, 27, 64, 125, 216, 343, 512]
lulesh_iterations: list[int] = [8, 16, 32, 64, 128, 256]
lulesh_sizes: list[int] = [16, 24, 32, 48, 64, 128]

train_lulesh_processes: list[int] = [8, 27, 64, 125, 216, 343]
train_lulesh_iterations: list[int] = [8, 16, 32, 64, 128]
train_lulesh_sizes: list[int] = [16, 24, 32, 48]

test_lulesh_processes: list[int] = [512, 729, 1000]
test_lulesh_iterations: list[int] = [256, 512, 1024]
test_lulesh_sizes: list[int] = [64, 96, 128]

# Extra-Pのオプション
modelerNames: list[str] = [
    # "refining", 
    "multi-parameter",
    "default", 
    # "basic --options poly_exponents=-1,0,1,2,3 log_exponents=0,1 force_combination_exponents=1 allow_negative_exponents=1"
    ]

modelerOption: str = """ --options \#spm=Basic \#spo=poly_exponents=-1,0,1,2,3,log_exponents=0,1,force_combination_exponents=1,allow_negative_exponents=True"""

list_modelName: list[str] = [
    "modelIp",
    "modelLog",
    "modelLinAndIp",
    "modelLinAndLog",
    "modelIpAndLin",
    "modelIpAndLog",
    "modelLogAndLin",
    "modelLogAndIp",
    "modelProcessDividedByProblemSize",
    "modelProblemSizeDividedByProcess",
    "modelInfiniteProductOfProblemSizeMultipliedByProcesses",
    "modelInfiniteProductOfProblemSizeDividedByProcesses",
    "modelLinearSumOf2elementCombination",
    "modelLinearSumOfElementCombinations",
    "modelLinearSumOf2elementCombinationWithSquared",
    "modelLinearSumOf2elementCombinationWithCubed",
    "modelSquareRootOfProcess",
    "modelSquareRootTimesOtherElems",
    "modelObeyOneParameter",
    "modelLin"
    # "modelBasicTree",
]
list_csvDir = [
    "./csv_files/lulesh_1st/",
    "./csv_files/lulesh_2nd/",
    "./csv_files/lulesh_3rd/",
]

DEBUG:__main__:hello


# benchmark_lulesh.html と同様の方法を再現する

expVar: list[str] = ["process", "iteration", "size"]
resVar: str = "Exclusive"

trainDF_lulesh: pd.DataFrame = ret_averaged_rawDF_lulesh(
    list_process=train_lulesh_processes,
    list_iteration=train_lulesh_iterations,
    list_size=train_lulesh_sizes,
    list_csvDir=list_csvDir,
    resVar=resVar,
)

testDF_lulesh: pd.DataFrame = ret_averaged_rawDF_lulesh(
    list_process=test_lulesh_processes[-1:],
    list_iteration=test_lulesh_iterations[-1:],
    list_size=test_lulesh_sizes[-1:],
    list_csvDir=list_csvDir,
    resVar=resVar,
)

trainDF_lulesh = trainDF_lulesh.reset_index()
testDF_lulesh = testDF_lulesh.reset_index()

functionNames_lulesh: list[str] = sorted(list(set(trainDF_lulesh["Name"])))

dict_symbols_lulesh = {}
for elem in expVar:
    dict_symbols_lulesh[elem] = symbols(elem, real=True)

target_env = [
    (dict_symbols_lulesh["size"], test_lulesh_sizes[-1]),
    (dict_symbols_lulesh["iteration"], test_lulesh_iterations[-1]),
    (dict_symbols_lulesh["process"], test_lulesh_processes[-1]),
]

modelerName: str = "multi-parameter"

resVar_all: str = "Exclusive"
resVar_perCall: str = "ExclusivePerCall"
resVar_call: str = "#Call"

dict_models_lulesh_all = {}
dict_models_lulesh_perCall = {}
dict_models_lulesh_call = {}

# モデル構築
for functionName in functionNames_lulesh:

    trainDF_lulesh_perFunc: pd.DataFrame = trainDF_lulesh[
        trainDF_lulesh["Name"] == functionName
    ]
    testDF_lulesh_perFUnc: pd.DataFrame = testDF_lulesh[
        testDF_lulesh["Name"] == functionName
    ]

    str_ExtraP_model_all: str = get_ExtraP_model(
        inputDF_perFunc=trainDF_lulesh_perFunc,
        expVar=expVar,
        resVar=resVar_all,
        functionName=functionName,
        benchmarkName="lulesh",
        modelerName=modelerName,
        modelerOption=modelerOption,
        dict_symbols=dict_symbols_lulesh,
    )
    dict_models_lulesh_all[functionName] = str_ExtraP_model_all

    str_ExtraP_model_perCall: str = get_ExtraP_model(
        inputDF_perFunc=trainDF_lulesh_perFunc,
        expVar=expVar,
        resVar=resVar_perCall,
        functionName=functionName,
        benchmarkName="lulesh",
        modelerName=modelerName,
        modelerOption=modelerOption,
        dict_symbols=dict_symbols_lulesh,
    )
    dict_models_lulesh_perCall[functionName] = str_ExtraP_model_perCall

    str_ExtraP_model_call: str = get_ExtraP_model(
        inputDF_perFunc=trainDF_lulesh_perFunc,
        expVar=expVar,
        resVar=resVar_call,
        functionName=functionName,
        benchmarkName="lulesh",
        modelerName=modelerName,
        modelerOption=modelerOption,
        dict_symbols=dict_symbols_lulesh,
    )
    dict_models_lulesh_call[functionName] = str_ExtraP_model_call

_list_series: list[pd.Series] = []
# 予測
for functionName in functionNames_lulesh:
    # 全体
    _all: float = dict_models_lulesh_all[functionName].subs(target_env_lulesh).evalf()
    # １コール当たり
    _perCall: float = (
        dict_models_lulesh_perCall[functionName].subs(target_env_lulesh).evalf()
    )
    # コール回数
    _call: float = dict_models_lulesh_call[functionName].subs(target_env_lulesh).evalf()

    target_lulesh_DF: pd.DataFrame = testDF_lulesh[
        (testDF_lulesh["Name"] == functionName)
        & (testDF_lulesh["process"] == test_lulesh_processes[-1])
        & (testDF_lulesh["iteration"] == test_lulesh_iterations[-1])
        & (testDF_lulesh["size"] == test_lulesh_sizes[-1])
    ]
    target_lulesh_DF = target_lulesh_DF.reset_index()
    # 実測値
    _real_exclusive: float = target_lulesh_DF.loc[0][resVar_all]
    _real_call: float = target_lulesh_DF.loc[0][resVar_call]

    _tmp_series: pd.Series = pd.Series(dtype="object")
    _tmp_series["Name"] = functionName
    _tmp_series["all to all"] = _all
    _tmp_series["perCall to all"] = _perCall * _call
    _tmp_series["real exclusive"] = _real_exclusive
    _tmp_series["real call"] = _real_call

    _list_series.append(_tmp_series)

resultDF: pd.DataFrame = pd.DataFrame(data=_list_series)
resultDF = add_relativeErrorRateCol(
    inputDF=resultDF,
    real_colName="real exclusive",
    predicted_colName="all to all",
    targetColName="vs all to all",
)
resultDF = add_relativeErrorRateCol(
    inputDF=resultDF,
    real_colName="real exclusive",
    predicted_colName="perCall to all",
    targetColName="vs perCall to all",
)

print(resultDF.mean())

resultDF

# "vs all to all" より "vs perCall to all" のほうが精度がMAPEベースで良いことを確認

all to all           1.682026e+02
perCall to all       1.061613e+02
real exclusive       1.453734e+02
real call            6.717865e+07
vs all to all        9.842000e+01
vs perCall to all    7.618677e+01
dtype: float64

/tmp/ipykernel_121/2418870204.py:123: FutureWarning:

The default value of numeric_only in DataFrame.mean is deprecated. In a future version, it will default to False. In addition, specifying 'numeric_only=None' is deprecated. Select only valid columns or specify the value of numeric_only to silence this warning.


resultDF_better_than_all_to_all: pd.DataFrame = resultDF[
    resultDF["vs all to all"] > resultDF["vs perCall to all"]
]
resultDF_better_than_all_to_all


floatWeightedMAPE_all: float = calcWeightedMAPEscore(
    inputDF=resultDF,
    inputColumnDict={
        "funcName": "Name",
        "call": "real call",
        "MAPE": "vs all to all",
    },
)
floatWeightedMAPE_perCall: float = calcWeightedMAPEscore(
    inputDF=resultDF,
    inputColumnDict={
        "funcName": "Name",
        "call": "real call",
        "MAPE": "vs perCall to all",
    },
)
pd.Series(
    {
        "weighted all to all": floatWeightedMAPE_all,
        "weighted perCall to all": floatWeightedMAPE_perCall,
    }
)

weighted all to all        5.98776005636523
weighted perCall to all    1.73978719275924
dtype: object


functionNames_lulesh_better_than_all_to_all: list[str] = sorted(
    list(set(resultDF_better_than_all_to_all["Name"]))
)
functionNames_lulesh_better_than_all_to_all

['MPI_Barrier()',
 'MPI_Comm_rank()',
 'MPI_Init()',
 'MPI_Isend()',
 'MPI_Wait()',
 'MPI_Waitall()',
 'Real_t_CalcElemVolume(const',
 'void_CalcKinematicsForElems(Domain',
 'void_CommMonoQ(Domain',
 'void_CommSyncPosVel(Domain',
 'void_Domain::CreateRegionIndexSets(Int_t_Int_t)',
 'void_Domain::SetupSymmetryPlanes(Int_t)',
 'void_ParseCommandLineOptions(int_char_**_Int_t_cmdLineOpts']


for functionName in functionNames_lulesh_better_than_all_to_all:
    resultDF_better_than_all_to_all_perFunc: pd.DataFrame = (
        resultDF_better_than_all_to_all[
            resultDF_better_than_all_to_all["Name"] == functionName
        ].reset_index()
    )
    try:
        _mape_all: float = resultDF_better_than_all_to_all_perFunc.loc[0][
            "vs all to all"
        ]
    except:
        _mape_all: float = -1.0
    try:
        _mape_perCall: float = resultDF_better_than_all_to_all_perFunc.loc[0][
            "vs perCall to all"
        ]
    except:
        _mape_perCall: float = -1.0
    print(f"\\item[{functionName}]\\mbox\\\\")
    print("\\begin{description}")
    print(f"\\item[直接方式({_mape_all})]")
    print("\\begin{dmath}")
    print(f"{sympy.latex(dict_models_lulesh_all[functionName])}")
    print("\\end{dmath}")
    print(f"\\item[分離方式({_mape_perCall})]")
    print("\\begin{dmath}")
    print(
        f"({sympy.latex(dict_models_lulesh_perCall[functionName])}) * ({sympy.latex(dict_models_lulesh_call[functionName])})"
    )
    print("\\end{dmath}")
    print("\\end{description}")
    print("\n")

\item[MPI_Barrier()]\mbox\\
\begin{description}
\item[直接方式(97.8704420289855)]
\begin{dmath}
0.0008163305555555552
\end{dmath}
\item[分離方式(27.1500185343787)]
\begin{dmath}
(2.5894115476979053 \cdot 10^{-8} iteration^{2} + 0.0007738781980994149) * (0.9999999999999989)
\end{dmath}
\end{description}


\item[MPI_Comm_rank()]\mbox\\
\begin{description}
\item[直接方式(84.1660497503293)]
\begin{dmath}
1.2634769214973346 \cdot 10^{-7} iteration size + 6.461875508750511 \cdot 10^{-8} size + 6.028571428571336 \cdot 10^{-6}
\end{dmath}
\item[分離方式(23.2192566747993)]
\begin{dmath}
(8.210259548958283 \cdot 10^{-7} - \frac{9.167435109616658 \cdot 10^{-6}}{size}) * (9.000000000000002 iteration + 5.000000000000079)
\end{dmath}
\end{description}


\item[MPI_Init()]\mbox\\
\begin{description}
\item[直接方式(19.8776634163958)]
\begin{dmath}
0.49301944444444473
\end{dmath}
\item[分離方式(19.4948537378115)]
\begin{dmath}
(0.49537500000000034) * (0.9999999999999989)
\end{dmath}
\end{description}


\item[MPI_Isend()]\mbox\\
\begin{description}
\item[直接方式(254.747116498683)]
\begin{dmath}
1.3993860867527394 \cdot 10^{-9} iteration size^{3} + 7.979056435526728 \cdot 10^{-8} size^{3} + 0.00249501250422031
\end{dmath}
\item[分離方式(27.6620996622256)]
\begin{dmath}
(1.5055130336533512 \cdot 10^{-11} size^{3} + 2.9729002280179477 \cdot 10^{-6} + \frac{3.976455855019544 \cdot 10^{-9} size^{3}}{iteration}) * (26.12330995183692 iteration + 14.478159722222374)
\end{dmath}
\end{description}


\item[MPI_Wait()]\mbox\\
\begin{description}
\item[直接方式(66.6930247500683)]
\begin{dmath}
3.329256869707227 \cdot 10^{-8} iteration size^{3} + 0.005503515999093116
\end{dmath}
\item[分離方式(43.0131480032898)]
\begin{dmath}
(1.0823000294871803 \cdot 10^{-9} size^{3} + 8.55947827071325 \cdot 10^{-6} + \frac{6.663734395402893 \cdot 10^{-9} size^{3}}{iteration}) * (26.12330995183692 iteration + 14.478159722222374)
\end{dmath}
\end{description}


\item[MPI_Waitall()]\mbox\\
\begin{description}
\item[直接方式(1525.96782981038)]
\begin{dmath}
8.713335972591822 \cdot 10^{-10} iteration^{2} size^{3} + 1.6546795016402215 \cdot 10^{-6} size^{3} + 0.015882696300209494
\end{dmath}
\item[分離方式(109.768192128864)]
\begin{dmath}
(3.831994679457664 \cdot 10^{-8} size^{3} + 0.0002249575098478947) * (2.9999999999999987 iteration + 1.0000000000000082)
\end{dmath}
\end{description}


\item[Real_t_CalcElemVolume(const]\mbox\\
\begin{description}
\item[直接方式(5.97939380831707)]
\begin{dmath}
1.562338241151975 \cdot 10^{-7} iteration size^{3} - 0.00013220805704912162 iteration + 0.004302233217541518
\end{dmath}
\item[分離方式(1.73677312221611)]
\begin{dmath}
(1.4244067448465714 \cdot 10^{-7} + \frac{2.841433578567945 \cdot 10^{-7}}{size}) * (1.000002222849363 iteration size^{3} + 0.9999521780095053 size^{3} - 1.0172315458026697)
\end{dmath}
\end{description}


\item[void_CalcKinematicsForElems(Domain]\mbox\\
\begin{description}
\item[直接方式(1.64212291281154)]
\begin{dmath}
2.0733238239747908 \cdot 10^{-7} iteration size^{3} + 0.00017105421681036588 process + 0.005592745678458262
\end{dmath}
\item[分離方式(1.00255932501224)]
\begin{dmath}
(2.0867997246578587 \cdot 10^{-7} size^{3} + 0.00017482559351284662) * (1.0 iteration + 5.681593783157123 \cdot 10^{-16})
\end{dmath}
\end{description}


\item[void_CommMonoQ(Domain]\mbox\\
\begin{description}
\item[直接方式(265.176604660229)]
\begin{dmath}
2.446998149980512 \cdot 10^{-9} iteration size^{3} + 3.5138071062346432 \cdot 10^{-6} iteration + 5.645486111246788 \cdot 10^{-5}
\end{dmath}
\item[分離方式(103.106381549652)]
\begin{dmath}
(1.1375226076740821 \cdot 10^{-10} iteration size^{2} + 5.798011853623222 \cdot 10^{-8} size^{2} - 2.2090209387745293 \cdot 10^{-6}) * (1.0 iteration + 5.681593783157123 \cdot 10^{-16})
\end{dmath}
\end{description}


\item[void_CommSyncPosVel(Domain]\mbox\\
\begin{description}
\item[直接方式(90.3228039083456)]
\begin{dmath}
1.0136090651622634 \cdot 10^{-8} iteration size^{3} + 0.0007645633385803624
\end{dmath}
\item[分離方式(34.8657736605796)]
\begin{dmath}
(\frac{1.1227901165638597 \cdot 10^{-6} iteration}{process} + 7.178827915605421 \cdot 10^{-9} size^{3} + 8.614385230642934 \cdot 10^{-6} - \frac{0.001333921557274115}{process}) * (1.0 iteration + 5.681593783157123 \cdot 10^{-16})
\end{dmath}
\end{description}


\item[void_Domain::CreateRegionIndexSets(Int_t_Int_t)]\mbox\\
\begin{description}
\item[直接方式(66.6735950377721)]
\begin{dmath}
4.437481108533314 \cdot 10^{-7} size^{2} - 4.940903785569108 \cdot 10^{-5} - \frac{0.00023893521592081838}{process}
\end{dmath}
\item[分離方式(65.1575089242126)]
\begin{dmath}
(4.6436058976512 \cdot 10^{-7} size^{2} - 5.86219300780953 \cdot 10^{-5} - \frac{0.00025557287968530976}{process}) * (0.9999999999999989)
\end{dmath}
\end{description}


\item[void_Domain::SetupSymmetryPlanes(Int_t)]\mbox\\
\begin{description}
\item[直接方式(36.1070982228629)]
\begin{dmath}
2.3353654107502677 \cdot 10^{-9} size^{2} + 3.4641997281972096 \cdot 10^{-7}
\end{dmath}
\item[分離方式(35.6061343466214)]
\begin{dmath}
(2.326003005061153 \cdot 10^{-9} size^{2} + 3.5770687473640016 \cdot 10^{-7}) * (0.9999999999999989)
\end{dmath}
\end{description}


\item[void_ParseCommandLineOptions(int_char_**_Int_t_cmdLineOpts]\mbox\\
\begin{description}
\item[直接方式(32.5388418079097)]
\begin{dmath}
2.0852777777777786 \cdot 10^{-6}
\end{dmath}
\item[分離方式(30.9692796610169)]
\begin{dmath}
(2.060583333333334 \cdot 10^{-6}) * (0.9999999999999989)
\end{dmath}
\end{description}


# benchmark_lulesh.html と同様の方法を再現する(@Inclusive)

expVar: list[str] = ["process", "iteration", "size"]
resVar: str = "Inclusive"

trainDF_lulesh: pd.DataFrame = ret_averaged_rawDF_lulesh(
    list_process=train_lulesh_processes,
    list_iteration=train_lulesh_iterations,
    list_size=train_lulesh_sizes,
    list_csvDir=list_csvDir,
    resVar=resVar,
)

testDF_lulesh: pd.DataFrame = ret_averaged_rawDF_lulesh(
    list_process=test_lulesh_processes[-1:],
    list_iteration=test_lulesh_iterations[-1:],
    list_size=test_lulesh_sizes[-1:],
    list_csvDir=list_csvDir,
    resVar=resVar,
)

trainDF_lulesh = trainDF_lulesh.reset_index()
testDF_lulesh = testDF_lulesh.reset_index()

functionNames_lulesh: list[str] = sorted(list(set(trainDF_lulesh["Name"])))

dict_symbols_lulesh = {}
for elem in expVar:
    dict_symbols_lulesh[elem] = symbols(elem, real=True)

target_env = [
    (dict_symbols_lulesh["size"], test_lulesh_sizes[-1]),
    (dict_symbols_lulesh["iteration"], test_lulesh_iterations[-1]),
    (dict_symbols_lulesh["process"], test_lulesh_processes[-1]),
]

modelerName: str = "multi-parameter"

resVar_all: str = "Inclusive"
resVar_perCall: str = "InclusivePerCall"
resVar_call: str = "#Call"

dict_models_lulesh_all = {}
dict_models_lulesh_perCall = {}
dict_models_lulesh_call = {}

# モデル構築
for functionName in functionNames_lulesh:

    trainDF_lulesh_perFunc: pd.DataFrame = trainDF_lulesh[
        trainDF_lulesh["Name"] == functionName
    ]
    testDF_lulesh_perFUnc: pd.DataFrame = testDF_lulesh[
        testDF_lulesh["Name"] == functionName
    ]

    str_ExtraP_model_all: str = get_ExtraP_model(
        inputDF_perFunc=trainDF_lulesh_perFunc,
        expVar=expVar,
        resVar=resVar_all,
        functionName=functionName,
        benchmarkName="lulesh",
        modelerName=modelerName,
        modelerOption=modelerOption,
        dict_symbols=dict_symbols_lulesh,
    )
    dict_models_lulesh_all[functionName] = str_ExtraP_model_all

    str_ExtraP_model_perCall: str = get_ExtraP_model(
        inputDF_perFunc=trainDF_lulesh_perFunc,
        expVar=expVar,
        resVar=resVar_perCall,
        functionName=functionName,
        benchmarkName="lulesh",
        modelerName=modelerName,
        modelerOption=modelerOption,
        dict_symbols=dict_symbols_lulesh,
    )
    dict_models_lulesh_perCall[functionName] = str_ExtraP_model_perCall

    str_ExtraP_model_call: str = get_ExtraP_model(
        inputDF_perFunc=trainDF_lulesh_perFunc,
        expVar=expVar,
        resVar=resVar_call,
        functionName=functionName,
        benchmarkName="lulesh",
        modelerName=modelerName,
        modelerOption=modelerOption,
        dict_symbols=dict_symbols_lulesh,
    )
    dict_models_lulesh_call[functionName] = str_ExtraP_model_call

_list_series: list[pd.Series] = []
# 予測
for functionName in functionNames_lulesh:
    # 全体
    _all: float = dict_models_lulesh_all[functionName].subs(target_env_lulesh).evalf()
    # １コール当たり
    _perCall: float = (
        dict_models_lulesh_perCall[functionName].subs(target_env_lulesh).evalf()
    )
    # コール回数
    _call: float = dict_models_lulesh_call[functionName].subs(target_env_lulesh).evalf()

    target_lulesh_DF: pd.DataFrame = testDF_lulesh[
        (testDF_lulesh["Name"] == functionName)
        & (testDF_lulesh["process"] == test_lulesh_processes[-1])
        & (testDF_lulesh["iteration"] == test_lulesh_iterations[-1])
        & (testDF_lulesh["size"] == test_lulesh_sizes[-1])
    ]
    target_lulesh_DF = target_lulesh_DF.reset_index()
    # 実測値
    _real_inclusive: float = target_lulesh_DF.loc[0][resVar_all]
    _real_call: float = target_lulesh_DF.loc[0][resVar_call]

    _tmp_series: pd.Series = pd.Series(dtype="object")
    _tmp_series["Name"] = functionName
    _tmp_series["all to all"] = _all
    _tmp_series["perCall to all"] = _perCall * _call
    _tmp_series["real inclusive"] = _real_inclusive
    _tmp_series["real call"] = _real_call

    _list_series.append(_tmp_series)

resultDF: pd.DataFrame = pd.DataFrame(data=_list_series)
resultDF = add_relativeErrorRateCol(
    inputDF=resultDF,
    real_colName="real inclusive",
    predicted_colName="all to all",
    targetColName="vs all to all",
)
resultDF = add_relativeErrorRateCol(
    inputDF=resultDF,
    real_colName="real inclusive",
    predicted_colName="perCall to all",
    targetColName="vs perCall to all",
)

print(resultDF.mean())

resultDF

# "vs all to all" より "vs perCall to all" のほうが精度がMAPEベースで良いことを確認

all to all           3.853701e+02
perCall to all       2.939183e+02
real inclusive       3.516647e+02
real call            6.717865e+07
vs all to all        1.170237e+02
vs perCall to all    5.267761e+01
dtype: float64

/tmp/ipykernel_121/4237747298.py:123: FutureWarning:

The default value of numeric_only in DataFrame.mean is deprecated. In a future version, it will default to False. In addition, specifying 'numeric_only=None' is deprecated. Select only valid columns or specify the value of numeric_only to silence this warning.


resultDF_better_than_all_to_all: pd.DataFrame = resultDF[
    resultDF["vs all to all"] > resultDF["vs perCall to all"]
]
resultDF_better_than_all_to_all


floatWeightedMAPE_all: float = calcWeightedMAPEscore(
    inputDF=resultDF,
    inputColumnDict={
        "funcName": "Name",
        "call": "real call",
        "MAPE": "vs all to all",
    },
)
floatWeightedMAPE_perCall: float = calcWeightedMAPEscore(
    inputDF=resultDF,
    inputColumnDict={
        "funcName": "Name",
        "call": "real call",
        "MAPE": "vs perCall to all",
    },
)
pd.Series(
    {
        "weighted all to all": floatWeightedMAPE_all,
        "weighted perCall to all": floatWeightedMAPE_perCall,
    }
)

weighted all to all        6.52310064960460
weighted perCall to all    1.73885845725729
dtype: object


# HPCA2022と同様のグラフをExtra-Pを用いて作成

# 予測対象をループ
# 各予測対象に対して予測
# 総実行時間を予測
# 表にする

resVar_call = "#Call"

for elem_process in test_lulesh_processes:
    for elem_iteration in test_lulesh_iterations:
        for elem_size in test_lulesh_sizes:
            pass

print(
    f"elem_process={elem_process}, elem_iteration={elem_iteration}, elem_size={elem_size}"
)

resVar_all = "Exclusive"
resVar_perCall = f"{resVar_all}PerCall"

resVar_all = "Inclusive"
resVar_perCall = f"{resVar_all}PerCall"

elem_process=1000, elem_iteration=1024, elem_size=128


%reset -f

# jupyter_pwd = %pwd
# if jupyter_pwd == "/":
#     %cd /workspace


# ipynb形式のライブラリのインポート
%run ./lib/lib.ipynb

# 生データの入ったCSVファイルの保持されたディレクトリ名を格納している変数
csvDirPath = "./csv_files/"

# NPBのベンチマーク名のリスト
benchmarkNames = ["cg", "ep", "ft", "is", "lu", "mg"]

# LULESH ベンチマークプログラムのプロセス数・問題サイズ・イテレーション数
lulesh_processes: list[int] = [8, 27, 64, 125, 216, 343, 512]
lulesh_iterations: list[int] = [8, 16, 32, 64, 128, 256]
lulesh_sizes: list[int] = [16, 24, 32, 48, 64, 128]

# ipynb形式のライブラリノートを.py形式に変更したものをインポート
import lib
import lib.lab_lib
from lib.lab_lib import *

train_lulesh_processes: list[int] = [8, 27, 64, 125, 216, 343]
train_lulesh_iterations: list[int] = [8, 16, 32, 64, 128]
train_lulesh_sizes: list[int] = [16, 24, 32, 48]

test_lulesh_processes: list[int] = [512, 729, 1000]
test_lulesh_iterations: list[int] = [256, 512, 1024]
test_lulesh_sizes: list[int] = [64, 96, 128]

list_modelName: list[str] = [
    "modelIp",
    "modelLog",
    "modelLinAndIp",
    "modelLinAndLog",
    "modelIpAndLin",
    "modelIpAndLog",
    "modelLogAndLin",
    "modelLogAndIp",
    "modelProcessDividedByProblemSize",
    "modelProblemSizeDividedByProcess",
    "modelInfiniteProductOfProblemSizeMultipliedByProcesses",
    "modelInfiniteProductOfProblemSizeDividedByProcesses",
    "modelLinearSumOf2elementCombination",
    "modelLinearSumOfElementCombinations",
    "modelLinearSumOf2elementCombinationWithSquared",
    "modelLinearSumOf2elementCombinationWithCubed",
    "modelSquareRootOfProcess",
    "modelSquareRootTimesOtherElems",
    "modelObeyOneParameter",
    "modelLin"
    # "modelBasicTree",
]


# ipynb形式のライブラリのインポート
%run ./lib/lib.ipynb

# ipynb形式のライブラリノートを.py形式に変更したものをインポート
import lib
import lib.lab_lib
from lib.lab_lib import *

# 生データの入ったCSVファイルの保持されたディレクトリ名を格納している変数
csvDirPath = "./csv_files/"

# NPBのベンチマーク名のリスト
benchmarkNames = ["cg", "ep", "ft", "is", "lu", "mg"]

# NPBのプロセス数
npb_process: list[int] = [2, 4, 8, 16, 32, 64, 128, 256]
train_npb_process: list[int] = npb_process[:-1]
test_npb_process: list[int] = npb_process[-1:]
# NPBのCGの初期変数
cg_na: list[int] = [14000, 30000, 75000, 100000, 1500000]
cg_nonzer: list[int] = [11, 12, 13, 14, 15, 18, 21]
cg_niter: list[int] = [15, 30, 75, 90, 100]
cg_shift: list[int] = [20, 40, 60, 80, 110, 200]

train_cg_na: list[int] = cg_na[:-1]
train_cg_nonzer: list[int] = cg_nonzer[:-1]
train_cg_niter: list[int] = cg_niter[:-1]
train_cg_shift: list[int] = cg_shift[:-1]

test_cg_na: list[int] = cg_na[-1:]
test_cg_nonzer: list[int] = cg_nonzer[-1:]
test_cg_niter: list[int] = cg_niter[-1:]
test_cg_shift: list[int] = cg_shift[-1:]

# LULESH ベンチマークプログラムのプロセス数・問題サイズ・イテレーション数
lulesh_processes: list[int] = [8, 27, 64, 125, 216, 343, 512]
lulesh_iterations: list[int] = [8, 16, 32, 64, 128, 256]
lulesh_sizes: list[int] = [16, 24, 32, 48, 64, 128]

train_lulesh_processes: list[int] = [8, 27, 64, 125, 216, 343]
train_lulesh_iterations: list[int] = [8, 16, 32, 64, 128]
train_lulesh_sizes: list[int] = [16, 24, 32, 48]

test_lulesh_processes: list[int] = [512, 729, 1000]
test_lulesh_iterations: list[int] = [256, 512, 1024]
test_lulesh_sizes: list[int] = [64, 96, 128]

# Extra-Pのオプション
modelerNames: list[str] = [
    # "refining",
    "multi-parameter",
    "default",
    # "basic --options poly_exponents=-1,0,1,2,3 log_exponents=0,1 force_combination_exponents=1 allow_negative_exponents=1"
]

modelerOption: str = """ --options \#spm=Basic \#spo=poly_exponents=-1,0,1,2,3,log_exponents=0,1,force_combination_exponents=1,allow_negative_exponents=True"""

list_modelName: list[str] = [
    "modelIp",
    "modelLog",
    "modelLinAndIp",
    "modelLinAndLog",
    "modelIpAndLin",
    "modelIpAndLog",
    "modelLogAndLin",
    "modelLogAndIp",
    "modelProcessDividedByProblemSize",
    "modelProblemSizeDividedByProcess",
    "modelInfiniteProductOfProblemSizeMultipliedByProcesses",
    "modelInfiniteProductOfProblemSizeDividedByProcesses",
    "modelLinearSumOf2elementCombination",
    "modelLinearSumOfElementCombinations",
    "modelLinearSumOf2elementCombinationWithSquared",
    "modelLinearSumOf2elementCombinationWithCubed",
    "modelSquareRootOfProcess",
    "modelSquareRootTimesOtherElems",
    "modelObeyOneParameter",
    "modelLin"
    # "modelBasicTree",
]
list_csvDir = [
    "./csv_files/lulesh_1st/",
    "./csv_files/lulesh_2nd/",
    "./csv_files/lulesh_3rd/",
]

DEBUG:__main__:hello
DEBUG:__main__:hello


# モデル構築関数


def returnModelByExtraP(
    input_benchmarkName: str,
    input_expVar: list[str],
    input_resVar: str,
    input_trainDF: pd.DataFrame,
    input_testDF: pd.DataFrame,
    input_dict_symbols: dict[str, any],
    input_modelerName: str,
    input_modelerOption: str,
):
    functionNames: list[str] = sorted(list(set(input_trainDF["Name"])))

    resDict: dict[str, dict[str, any]] = {}
    dict_functionName_model_all: dict[str, any] = {}
    dict_functionName_model_perCall: dict[str, any] = {}

    resVar_all: str = input_resVar
    resVar_perCall: str = f"{resVar_all}PerCall"

    for functionName in functionNames:
        trainDF_perFunc: pd.DataFrame = input_trainDF[
            input_trainDF["Name"] == functionName
        ].reset_index()
        model_fromExtraP_all = get_ExtraP_model(
            benchmarkName=input_benchmarkName,
            inputDF_perFunc=trainDF_perFunc,
            expVar=input_expVar,
            resVar=resVar_all,
            functionName=functionName,
            modelerName=input_modelerName,
            modelerOption=input_modelerOption,
            dict_symbols=input_dict_symbols,
        )
        dict_functionName_model_all[functionName] = model_fromExtraP_all

        model_fromExtraP_perCall = get_ExtraP_model(
            benchmarkName=input_benchmarkName,
            inputDF_perFunc=trainDF_perFunc,
            expVar=input_expVar,
            resVar=resVar_perCall,
            functionName=functionName,
            modelerName=input_modelerName,
            modelerOption=input_modelerOption,
            dict_symbols=input_dict_symbols,
        )
        dict_functionName_model_perCall[functionName] = model_fromExtraP_perCall

    resDict["all"] = dict_functionName_model_all
    resDict["perCall"] = dict_functionName_model_perCall

    return resDict


# モデル構築

expVar: list[str] = ["process", "iteration", "size"]
resVar_ex: str = "Exclusive"
resVar_in: str = "Inclusive"


benchmarkName = "lulesh"

trainDF_lulesh_in: pd.DataFrame = ret_averaged_rawDF_lulesh(
    list_process=train_lulesh_processes,
    list_iteration=train_lulesh_iterations,
    list_size=train_lulesh_sizes,
    list_csvDir=list_csvDir,
    resVar=resVar_in,
)
testDF_lulesh_in: pd.DataFrame = ret_averaged_rawDF_lulesh(
    list_process=test_lulesh_processes,
    list_iteration=test_lulesh_iterations,
    list_size=test_lulesh_sizes,
    list_csvDir=list_csvDir,
    resVar=resVar_in,
)
trainDF_lulesh_in = trainDF_lulesh_in.reset_index()
testDF_lulesh_in = testDF_lulesh_in.reset_index()

trainDF_lulesh_ex: pd.DataFrame = ret_averaged_rawDF_lulesh(
    list_process=train_lulesh_processes,
    list_iteration=train_lulesh_iterations,
    list_size=train_lulesh_sizes,
    list_csvDir=list_csvDir,
    resVar=resVar_ex,
)
testDF_lulesh_ex: pd.DataFrame = ret_averaged_rawDF_lulesh(
    list_process=test_lulesh_processes,
    list_iteration=test_lulesh_iterations,
    list_size=test_lulesh_sizes,
    list_csvDir=list_csvDir,
    resVar=resVar_ex,
)
trainDF_lulesh_ex = trainDF_lulesh_ex.reset_index()
testDF_lulesh_ex = testDF_lulesh_ex.reset_index()

functionNames_lulesh: list[str] = sorted(list(set(trainDF_lulesh_in["Name"])))

dict_symbols_lulesh = {}
for elem in expVar:
    dict_symbols_lulesh[elem] = symbols(elem, real=True)

target_env = [
    (dict_symbols_lulesh["size"], test_lulesh_sizes[-1]),
    (dict_symbols_lulesh["iteration"], test_lulesh_iterations[-1]),
    (dict_symbols_lulesh["process"], test_lulesh_processes[-1]),
]

modelerName: str = "multi-parameter"

resVar_call: str = "#Call"


models_dict_ex: dict[str, any] = returnModelByExtraP(
    input_benchmarkName=benchmarkName,
    input_expVar=expVar,
    input_resVar=resVar_ex,
    input_trainDF=trainDF_lulesh_ex,
    input_testDF=testDF_lulesh_ex,
    input_dict_symbols=dict_symbols_lulesh,
    input_modelerName=modelerName,
    input_modelerOption=modelerOption,
)

models_dict_in: dict[str, any] = returnModelByExtraP(
    input_benchmarkName=benchmarkName,
    input_expVar=expVar,
    input_resVar=resVar_in,
    input_trainDF=trainDF_lulesh_in,
    input_testDF=testDF_lulesh_in,
    input_dict_symbols=dict_symbols_lulesh,
    input_modelerName=modelerName,
    input_modelerOption=modelerOption,
)

# コール回数のモデルを作成
resVar_call: str = f"#Call"
dict_functionName_model_call: dict[str, any] = {}

trainDF_lulesh_call: pd.DataFrame = ret_averaged_rawDF_lulesh(
    list_process=train_lulesh_processes,
    list_iteration=train_lulesh_iterations,
    list_size=train_lulesh_sizes,
    list_csvDir=list_csvDir,
    resVar=resVar_call,
)
testDF_lulesh_call: pd.DataFrame = ret_averaged_rawDF_lulesh(
    list_process=test_lulesh_processes,
    list_iteration=test_lulesh_iterations,
    list_size=test_lulesh_sizes,
    list_csvDir=list_csvDir,
    resVar=resVar_call,
)

functionNames_lulesh: list[str] = sorted(list(set(trainDF_lulesh_call["Name"])))
for functionName in functionNames_lulesh:
    trainDF_perFunc: pd.DataFrame = trainDF_lulesh_call[
        trainDF_lulesh_call["Name"] == functionName
    ].reset_index()

    model_fromExtraP_call = get_ExtraP_model(
        benchmarkName="lulesh",
        inputDF_perFunc=trainDF_perFunc,
        expVar=expVar,
        resVar=resVar_call,
        functionName=functionName,
        modelerName=modelerName,
        modelerOption=modelerOption,
        dict_symbols=dict_symbols_lulesh,
    )
    dict_functionName_model_call[functionName] = model_fromExtraP_call


# TODO
# テスト環境をループ
# meanでそれぞれの環境のデータを集計
_list_series_whole: list[pd.DataFrame] = []
for elem_process in test_lulesh_processes:
    for elem_iteration in test_lulesh_iterations:
        # コール回数の予測値を算出
        for elem_size in test_lulesh_sizes:
            resultDF_oneCase_ex: pd.DataFrame
            resultDF_oneCase_in: pd.DataFrame

            _list_series_perFuncs: list[pd.Series] = []
            for functionName in functionNames_lulesh:
                testDF_perFunc: pd.DataFrame = testDF_lulesh_call[
                    (testDF_lulesh_call["process"] == elem_process)
                    & (testDF_lulesh_call["iteration"] == elem_iteration)
                    & (testDF_lulesh_call["size"] == elem_size)
                    & (testDF_lulesh_call["Name"] == functionName)
                ].reset_index()

                testDF_perFunc_ex: pd.DataFrame = testDF_lulesh_ex[
                    (testDF_lulesh_ex["process"] == elem_process)
                    & (testDF_lulesh_ex["iteration"] == elem_iteration)
                    & (testDF_lulesh_ex["size"] == elem_size)
                    & (testDF_lulesh_ex["Name"] == functionName)
                ].reset_index()

                testDF_perFunc_in: pd.DataFrame = testDF_lulesh_ex[
                    (testDF_lulesh_in["process"] == elem_process)
                    & (testDF_lulesh_in["iteration"] == elem_iteration)
                    & (testDF_lulesh_in["size"] == elem_size)
                    & (testDF_lulesh_in["Name"] == functionName)
                ].reset_index()

                # 予測値の算出
                target_env_oneCase = [
                    (dict_symbols_lulesh["size"], elem_size),
                    (dict_symbols_lulesh["iteration"], elem_iteration),
                    (dict_symbols_lulesh["process"], elem_process),
                ]
                # _all
                _all_ex: float = (
                    models_dict_ex["all"][functionName].subs(target_env_oneCase).evalf()
                )
                _all_in: float = (
                    models_dict_in["all"][functionName].subs(target_env_oneCase).evalf()
                )
                # _perCall
                _perCall_ex: float = (
                    models_dict_ex["perCall"][functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )
                _perCall_in: float = (
                    models_dict_in["perCall"][functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )
                # _call
                _call: float = (
                    dict_functionName_model_call[functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )

                # 実測値の算出
                _real_ex: float = convertPprofTime(testDF_perFunc_ex.loc[0][resVar_ex])
                _real_in: float = convertPprofTime(testDF_perFunc_in.loc[0][resVar_in])

                _tmp_series: pd.Series = pd.Series(
                    {
                        "Name": functionName,
                        "process": elem_process,
                        "iteration": elem_iteration,
                        "size": elem_size,
                        "ex_directly": _all_ex,
                        "ex_indirectly": _perCall_ex * _call,
                        "ex_real": _real_ex,
                        "in_directly": _all_in,
                        "in_indirectly": _perCall_in * _call,
                        "in_real": _real_in,
                    }
                )
                _list_series_perFuncs.append(_tmp_series)
            _tmp_DF: pd.DataFrame = pd.DataFrame(data=_list_series_perFuncs)
            _tmp_DF = _tmp_DF.astype(
                {
                    "ex_directly": float,
                    "ex_indirectly": float,
                    "ex_real": float,
                    "in_directly": float,
                    "in_indirectly": float,
                    "in_real": float,
                }
            )
            _tmp_DF = add_relativeErrorRateCol(
                inputDF=_tmp_DF,
                real_colName="ex_real",
                predicted_colName="ex_directly",
                targetColName="vs directly(ex)",
            )
            _tmp_DF = add_relativeErrorRateCol(
                inputDF=_tmp_DF,
                real_colName="ex_real",
                predicted_colName="ex_indirectly",
                targetColName="vs indirectly(ex)",
            )
            _tmp_DF = add_relativeErrorRateCol(
                inputDF=_tmp_DF,
                real_colName="in_real",
                predicted_colName="in_directly",
                targetColName="vs directly(in)",
            )
            _tmp_DF = add_relativeErrorRateCol(
                inputDF=_tmp_DF,
                real_colName="in_real",
                predicted_colName="in_indirectly",
                targetColName="vs indirectly(in)",
            )
            _list_series_whole.append(_tmp_DF.mean(numeric_only=True))

resultDF_whole: pd.DataFrame = pd.DataFrame(data=_list_series_whole)
resultDF_whole


_tmp_DF


_tmp_DF.dtypes

Name                  object
process                int64
iteration              int64
size                   int64
ex_directly          float64
ex_indirectly        float64
ex_real              float64
in_directly          float64
in_indirectly        float64
in_real              float64
vs directly(ex)      float64
vs indirectly(ex)    float64
vs directly(in)      float64
vs indirectly(in)    float64
dtype: object


_tmp_DF.mean()

/tmp/ipykernel_121/1296138598.py:1: FutureWarning:

The default value of numeric_only in DataFrame.mean is deprecated. In a future version, it will default to False. In addition, specifying 'numeric_only=None' is deprecated. Select only valid columns or specify the value of numeric_only to silence this warning.

process              1.000000e+03
iteration            1.024000e+03
size                 1.280000e+02
ex_directly          1.682026e+02
ex_indirectly        1.005981e+04
ex_real              1.453734e-01
in_directly          3.853701e+02
in_indirectly        1.046825e+04
in_real              3.513761e+02
vs directly(ex)      1.541409e+05
vs indirectly(ex)    6.691295e+08
vs directly(in)      1.149511e+02
vs indirectly(in)    7.131300e+05
dtype: float64


print(f"{test_lulesh_processes[-1:]=}")
print(f"{test_lulesh_iterations[-1:]=}")
print(f"{test_lulesh_sizes[-1:]=}")

test_lulesh_processes[-1:]=[1000]
test_lulesh_iterations[-1:]=[1024]
test_lulesh_sizes[-1:]=[128]


expVar: list[str] = ["process", "iteration", "size"]
resVar: str = "Inclusive"

trainDF_lulesh: pd.DataFrame = ret_averaged_rawDF_lulesh(
    list_process=train_lulesh_processes,
    list_iteration=train_lulesh_iterations,
    list_size=train_lulesh_sizes,
    list_csvDir=list_csvDir,
    resVar=resVar,
)

testDF_lulesh: pd.DataFrame = ret_averaged_rawDF_lulesh(
    list_process=test_lulesh_processes,
    list_iteration=test_lulesh_iterations,
    list_size=test_lulesh_sizes,
    list_csvDir=list_csvDir,
    resVar=resVar,
)

trainDF_lulesh = trainDF_lulesh.reset_index()
testDF_lulesh = testDF_lulesh.reset_index()

functionNames_lulesh: list[str] = sorted(list(set(trainDF_lulesh["Name"])))

dict_symbols_lulesh = {}
for elem in expVar:
    dict_symbols_lulesh[elem] = symbols(elem, real=True)

target_env = [
    (dict_symbols_lulesh["size"], test_lulesh_sizes[-1]),
    (dict_symbols_lulesh["iteration"], test_lulesh_iterations[-1]),
    (dict_symbols_lulesh["process"], test_lulesh_processes[-1]),
]
target_env_lulesh = [
    (dict_symbols_lulesh["size"], test_lulesh_sizes[-1]),
    (dict_symbols_lulesh["iteration"], test_lulesh_iterations[-1]),
    (dict_symbols_lulesh["process"], test_lulesh_processes[-1]),
]

modelerName: str = "multi-parameter"

resVar_all: str = "Inclusive"
resVar_perCall: str = "InclusivePerCall"
resVar_call: str = "#Call"

dict_models_lulesh_all = {}
dict_models_lulesh_perCall = {}
dict_models_lulesh_call = {}

# モデル構築
for functionName in functionNames_lulesh:

    trainDF_lulesh_perFunc: pd.DataFrame = trainDF_lulesh[
        trainDF_lulesh["Name"] == functionName
    ]
    testDF_lulesh_perFUnc: pd.DataFrame = testDF_lulesh[
        testDF_lulesh["Name"] == functionName
    ]

    str_ExtraP_model_all: str = get_ExtraP_model(
        inputDF_perFunc=trainDF_lulesh_perFunc,
        expVar=expVar,
        resVar=resVar_all,
        functionName=functionName,
        benchmarkName="lulesh",
        modelerName=modelerName,
        modelerOption=modelerOption,
        dict_symbols=dict_symbols_lulesh,
    )
    dict_models_lulesh_all[functionName] = str_ExtraP_model_all

    str_ExtraP_model_perCall: str = get_ExtraP_model(
        inputDF_perFunc=trainDF_lulesh_perFunc,
        expVar=expVar,
        resVar=resVar_perCall,
        functionName=functionName,
        benchmarkName="lulesh",
        modelerName=modelerName,
        modelerOption=modelerOption,
        dict_symbols=dict_symbols_lulesh,
    )
    dict_models_lulesh_perCall[functionName] = str_ExtraP_model_perCall

    str_ExtraP_model_call: str = get_ExtraP_model(
        inputDF_perFunc=trainDF_lulesh_perFunc,
        expVar=expVar,
        resVar=resVar_call,
        functionName=functionName,
        benchmarkName="lulesh",
        modelerName=modelerName,
        modelerOption=modelerOption,
        dict_symbols=dict_symbols_lulesh,
    )
    dict_models_lulesh_call[functionName] = str_ExtraP_model_call

_list_series: list[pd.Series] = []
# 予測
for functionName in functionNames_lulesh:
    # 全体
    _all: float = dict_models_lulesh_all[functionName].subs(target_env_lulesh).evalf()
    # １コール当たり
    _perCall: float = (
        dict_models_lulesh_perCall[functionName].subs(target_env_lulesh).evalf()
    )
    # コール回数
    _call: float = dict_models_lulesh_call[functionName].subs(target_env_lulesh).evalf()

    target_lulesh_DF: pd.DataFrame = testDF_lulesh[
        (testDF_lulesh["Name"] == functionName)
        & (testDF_lulesh["process"] == test_lulesh_processes[-1])
        & (testDF_lulesh["iteration"] == test_lulesh_iterations[-1])
        & (testDF_lulesh["size"] == test_lulesh_sizes[-1])
    ]
    target_lulesh_DF = target_lulesh_DF.reset_index()
    # 実測値
    _real_inclusive: float = target_lulesh_DF.loc[0][resVar_all]
    _real_call: float = target_lulesh_DF.loc[0][resVar_call]

    _tmp_series: pd.Series = pd.Series(dtype="object")
    _tmp_series["Name"] = functionName
    _tmp_series["all to all"] = _all
    _tmp_series["perCall to all"] = _perCall * _call
    _tmp_series["real inclusive"] = _real_inclusive
    _tmp_series["real call"] = _real_call

    _list_series.append(_tmp_series)

resultDF: pd.DataFrame = pd.DataFrame(data=_list_series)
resultDF = add_relativeErrorRateCol(
    inputDF=resultDF,
    real_colName="real inclusive",
    predicted_colName="all to all",
    targetColName="vs all to all",
)
resultDF = add_relativeErrorRateCol(
    inputDF=resultDF,
    real_colName="real inclusive",
    predicted_colName="perCall to all",
    targetColName="vs perCall to all",
)

print(resultDF.mean())

resultDF

all to all           3.853701e+02
perCall to all       2.939183e+02
real inclusive       3.516647e+02
real call            6.717865e+07
vs all to all        1.170237e+02
vs perCall to all    5.267761e+01
dtype: float64

/tmp/ipykernel_121/2635962801.py:142: FutureWarning:

The default value of numeric_only in DataFrame.mean is deprecated. In a future version, it will default to False. In addition, specifying 'numeric_only=None' is deprecated. Select only valid columns or specify the value of numeric_only to silence this warning.


_list_series: list[pd.Series] = []
for elem_process in test_lulesh_processes:
    for elem_iteration in test_lulesh_iterations:
        for elem_size in test_lulesh_sizes:
            _list_series_oneCase: list[pd.Series] = []
            for functionName in functionNames_lulesh:
                target_env_oneCase = [
                    (dict_symbols_lulesh["size"], elem_size),
                    (dict_symbols_lulesh["iteration"], elem_iteration),
                    (dict_symbols_lulesh["process"], elem_process),
                ]
                # 全体
                _all: float = (
                    dict_models_lulesh_all[functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )
                # １コール当たり
                _perCall: float = (
                    dict_models_lulesh_perCall[functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )
                # コール回数
                _call: float = (
                    dict_models_lulesh_call[functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )

                target_lulesh_DF: pd.DataFrame = testDF_lulesh[
                    (testDF_lulesh["Name"] == functionName)
                    & (testDF_lulesh["process"] == elem_process)
                    & (testDF_lulesh["iteration"] == elem_iteration)
                    & (testDF_lulesh["size"] == elem_size)
                ]
                target_lulesh_DF = target_lulesh_DF.reset_index()
                # 実測値
                _real_inclusive: float = target_lulesh_DF.loc[0][resVar_all]
                _real_call: float = target_lulesh_DF.loc[0][resVar_call]

                _tmp_series: pd.Series = pd.Series(dtype="object")
                _tmp_series["Name"] = functionName
                _tmp_series["all to all"] = _all
                _tmp_series["perCall to all"] = _perCall * _call
                _tmp_series["real inclusive"] = _real_inclusive
                _tmp_series["real call"] = _real_call

                _list_series_oneCase.append(_tmp_series)

            resultDF: pd.DataFrame = pd.DataFrame(data=_list_series_oneCase)
            resultDF = add_relativeErrorRateCol(
                inputDF=resultDF,
                real_colName="real inclusive",
                predicted_colName="all to all",
                targetColName="vs all to all",
            )
            resultDF = add_relativeErrorRateCol(
                inputDF=resultDF,
                real_colName="real inclusive",
                predicted_colName="perCall to all",
                targetColName="vs perCall to all",
            )
            resultDF["process"] = elem_process
            resultDF["iteration"] = elem_iteration
            resultDF["size"] = elem_size

            resultDF["weighted all"] = calcWeightedMAPEscore(
                inputDF=resultDF,
                inputColumnDict={
                    "funcName": "Name",
                    "call": "real inclusive",
                    "MAPE": "vs all to all",
                },
            )
            resultDF["weighted perCall"] = calcWeightedMAPEscore(
                inputDF=resultDF,
                inputColumnDict={
                    "funcName": "Name",
                    "call": "real inclusive",
                    "MAPE": "vs perCall to all",
                },
            )

            resultDF = resultDF.astype(
                {
                    "all to all": float,
                    "perCall to all": float,
                    "real inclusive": float,
                    "real call": float,
                    "vs all to all": float,
                    "vs perCall to all": float,
                    "weighted perCall": float,
                    "weighted all": float,
                }
            )

            _list_series.append(resultDF.mean(numeric_only=True))

resultDF_in: pd.DataFrame = pd.DataFrame(data=_list_series)
resultDF_in


_list_series: list[pd.Series] = []
for elem_process in train_lulesh_processes:
    for elem_iteration in train_lulesh_iterations:
        for elem_size in train_lulesh_sizes:
            for functionName in functionNames_lulesh:
                target_env_oneCase = [
                    (dict_symbols_lulesh["size"], elem_size),
                    (dict_symbols_lulesh["iteration"], elem_iteration),
                    (dict_symbols_lulesh["process"], elem_process),
                ]

                # 全体
                _all: float = (
                    dict_models_lulesh_all[functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )
                # 1コール当たり
                _perCall: float = (
                    dict_models_lulesh_perCall[functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )
                # コール回数
                _call: float = (
                    dict_models_lulesh_call[functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )

                target_lulesh_DF: pd.DataFrame = trainDF_lulesh[
                    (trainDF_lulesh["Name"] == functionName)
                    & (trainDF_lulesh["process"] == elem_process)
                    & (trainDF_lulesh["iteration"] == elem_iteration)
                    & (trainDF_lulesh["size"] == elem_size)
                ]

                _real_in: float = target_lulesh_DF.reset_index().loc[0]["Inclusive"]
                _real_call: float = target_lulesh_DF.reset_index().loc[0]["#Call"]

                _tmp_series: pd.Series = pd.Series(
                    {
                        "Name": functionName,
                        "process": elem_process,
                        "iteration": elem_iteration,
                        "size": elem_size,
                        "_all": _all,
                        "_perCall": _perCall * _call,
                        "_call": _call,
                        "_real_call": _real_call,
                        "_real_in": _real_in,
                    }
                )

                _list_series.append(_tmp_series)


resultDF_in_fitness: pd.DataFrame = pd.DataFrame(_list_series)
resultDF_in_fitness = add_relativeErrorRateCol(
    inputDF=resultDF_in_fitness,
    real_colName="_real_in",
    predicted_colName="_all",
    targetColName="vs _all",
)
resultDF_in_fitness = add_relativeErrorRateCol(
    inputDF=resultDF_in_fitness,
    real_colName="_real_in",
    predicted_colName="_perCall",
    targetColName="vs _perCall",
)
resultDF_in_fitness


resultDF_in_fitness.mean()

/tmp/ipykernel_121/3270339140.py:1: FutureWarning:

The default value of numeric_only in DataFrame.mean is deprecated. In a future version, it will default to False. In addition, specifying 'numeric_only=None' is deprecated. Select only valid columns or specify the value of numeric_only to silence this warning.

process          130.500000
iteration         49.600000
size              30.000000
_all               0.567709
_perCall           0.520875
_call          63915.372350
_real_call     63915.372350
_real_in           0.567709
vs _all          145.280166
vs _perCall      146.162156
dtype: float64


# exclusive, モデル適合度

_list_series: list[pd.Series] = []
for functionName in functionNames_lulesh:
    _list_series_oneCase: list[pd.Series] = []
    for elem_process in train_lulesh_processes:
        for elem_iteration in train_lulesh_iterations:
            for elem_size in train_lulesh_sizes:
                target_env_oneCase = [
                    (dict_symbols_lulesh["size"], elem_size),
                    (dict_symbols_lulesh["iteration"], elem_iteration),
                    (dict_symbols_lulesh["process"], elem_process),
                ]
                # 全体
                _all: float = (
                    dict_models_lulesh_all[functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )
                # １コール当たり
                _perCall: float = (
                    dict_models_lulesh_perCall[functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )
                # コール回数
                _call: float = (
                    dict_models_lulesh_call[functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )

                target_lulesh_DF: pd.DataFrame = trainDF_lulesh[
                    (trainDF_lulesh["Name"] == functionName)
                    & (trainDF_lulesh["process"] == elem_process)
                    & (trainDF_lulesh["iteration"] == elem_iteration)
                    & (trainDF_lulesh["size"] == elem_size)
                ]
                target_lulesh_DF = target_lulesh_DF.reset_index()
                # 実測値
                _real_exclusive: float = target_lulesh_DF.loc[0][resVar_all]
                _real_call: float = target_lulesh_DF.loc[0][resVar_call]

                _tmp_series: pd.Series = pd.Series(dtype="object")
                _tmp_series["Name"] = functionName
                _tmp_series["all to all"] = _all
                _tmp_series["perCall to all"] = _perCall * _call
                _tmp_series["real inclusive"] = _real_exclusive
                _tmp_series["real call"] = _real_call

                _list_series_oneCase.append(_tmp_series)

    resultDF: pd.DataFrame = pd.DataFrame(data=_list_series_oneCase)
    resultDF = add_relativeErrorRateCol(
        inputDF=resultDF,
        real_colName="real inclusive",
        predicted_colName="all to all",
        targetColName="vs all to all",
    )
    resultDF = add_relativeErrorRateCol(
        inputDF=resultDF,
        real_colName="real inclusive",
        predicted_colName="perCall to all",
        targetColName="vs perCall to all",
    )

    resultDF = resultDF.astype(
        {
            "all to all": float,
            "perCall to all": float,
            "real inclusive": float,
            "real call": float,
            "vs all to all": float,
            "vs perCall to all": float,
        }
    )

    _tmp_list_series: pd.Series = resultDF.mean(numeric_only=True)
    _tmp_list_series["Name"] = functionName
    _list_series.append(_tmp_list_series)

resultDF_in_fitness: pd.DataFrame = pd.DataFrame(_list_series)
resultDF_in_fitness


# Exclusive


expVar: list[str] = ["process", "iteration", "size"]
resVar: str = "Exclusive"

trainDF_lulesh: pd.DataFrame = ret_averaged_rawDF_lulesh(
    list_process=train_lulesh_processes,
    list_iteration=train_lulesh_iterations,
    list_size=train_lulesh_sizes,
    list_csvDir=list_csvDir,
    resVar=resVar,
)

testDF_lulesh: pd.DataFrame = ret_averaged_rawDF_lulesh(
    list_process=test_lulesh_processes,
    list_iteration=test_lulesh_iterations,
    list_size=test_lulesh_sizes,
    list_csvDir=list_csvDir,
    resVar=resVar,
)

trainDF_lulesh = trainDF_lulesh.reset_index()
testDF_lulesh = testDF_lulesh.reset_index()

functionNames_lulesh: list[str] = sorted(list(set(trainDF_lulesh["Name"])))

dict_symbols_lulesh = {}
for elem in expVar:
    dict_symbols_lulesh[elem] = symbols(elem, real=True)

target_env = [
    (dict_symbols_lulesh["size"], test_lulesh_sizes[-1]),
    (dict_symbols_lulesh["iteration"], test_lulesh_iterations[-1]),
    (dict_symbols_lulesh["process"], test_lulesh_processes[-1]),
]
target_env_lulesh = [
    (dict_symbols_lulesh["size"], test_lulesh_sizes[-1]),
    (dict_symbols_lulesh["iteration"], test_lulesh_iterations[-1]),
    (dict_symbols_lulesh["process"], test_lulesh_processes[-1]),
]

modelerName: str = "multi-parameter"

resVar_all: str = "Exclusive"
resVar_perCall: str = "ExclusivePerCall"
resVar_call: str = "#Call"

dict_models_lulesh_all = {}
dict_models_lulesh_perCall = {}
dict_models_lulesh_call = {}

# モデル構築
for functionName in functionNames_lulesh:

    trainDF_lulesh_perFunc: pd.DataFrame = trainDF_lulesh[
        trainDF_lulesh["Name"] == functionName
    ]
    testDF_lulesh_perFUnc: pd.DataFrame = testDF_lulesh[
        testDF_lulesh["Name"] == functionName
    ]

    str_ExtraP_model_all: str = get_ExtraP_model(
        inputDF_perFunc=trainDF_lulesh_perFunc,
        expVar=expVar,
        resVar=resVar_all,
        functionName=functionName,
        benchmarkName="lulesh",
        modelerName=modelerName,
        modelerOption=modelerOption,
        dict_symbols=dict_symbols_lulesh,
    )
    dict_models_lulesh_all[functionName] = str_ExtraP_model_all

    str_ExtraP_model_perCall: str = get_ExtraP_model(
        inputDF_perFunc=trainDF_lulesh_perFunc,
        expVar=expVar,
        resVar=resVar_perCall,
        functionName=functionName,
        benchmarkName="lulesh",
        modelerName=modelerName,
        modelerOption=modelerOption,
        dict_symbols=dict_symbols_lulesh,
    )
    dict_models_lulesh_perCall[functionName] = str_ExtraP_model_perCall

    str_ExtraP_model_call: str = get_ExtraP_model(
        inputDF_perFunc=trainDF_lulesh_perFunc,
        expVar=expVar,
        resVar=resVar_call,
        functionName=functionName,
        benchmarkName="lulesh",
        modelerName=modelerName,
        modelerOption=modelerOption,
        dict_symbols=dict_symbols_lulesh,
    )
    dict_models_lulesh_call[functionName] = str_ExtraP_model_call

_list_series: list[pd.Series] = []
# 予測
for functionName in functionNames_lulesh:
    # 全体
    _all: float = dict_models_lulesh_all[functionName].subs(target_env_lulesh).evalf()
    # １コール当たり
    _perCall: float = (
        dict_models_lulesh_perCall[functionName].subs(target_env_lulesh).evalf()
    )
    # コール回数
    _call: float = dict_models_lulesh_call[functionName].subs(target_env_lulesh).evalf()

    target_lulesh_DF: pd.DataFrame = testDF_lulesh[
        (testDF_lulesh["Name"] == functionName)
        & (testDF_lulesh["process"] == test_lulesh_processes[-1])
        & (testDF_lulesh["iteration"] == test_lulesh_iterations[-1])
        & (testDF_lulesh["size"] == test_lulesh_sizes[-1])
    ]
    target_lulesh_DF = target_lulesh_DF.reset_index()
    # 実測値
    _real_exclusive: float = target_lulesh_DF.loc[0][resVar_all]
    _real_call: float = target_lulesh_DF.loc[0][resVar_call]

    _tmp_series: pd.Series = pd.Series(dtype="object")
    _tmp_series["Name"] = functionName
    _tmp_series["all to all"] = _all
    _tmp_series["perCall to all"] = _perCall * _call
    _tmp_series["real exclusive"] = _real_exclusive
    _tmp_series["real call"] = _real_call

    _list_series.append(_tmp_series)

resultDF: pd.DataFrame = pd.DataFrame(data=_list_series)
resultDF = add_relativeErrorRateCol(
    inputDF=resultDF,
    real_colName="real exclusive",
    predicted_colName="all to all",
    targetColName="vs all to all",
)
resultDF = add_relativeErrorRateCol(
    inputDF=resultDF,
    real_colName="real exclusive",
    predicted_colName="perCall to all",
    targetColName="vs perCall to all",
)

print(resultDF.mean())

resultDF

all to all           1.682026e+02
perCall to all       1.061613e+02
real exclusive       1.453734e+02
real call            6.717865e+07
vs all to all        9.842000e+01
vs perCall to all    7.618677e+01
dtype: float64

/tmp/ipykernel_121/3098613647.py:142: FutureWarning:

The default value of numeric_only in DataFrame.mean is deprecated. In a future version, it will default to False. In addition, specifying 'numeric_only=None' is deprecated. Select only valid columns or specify the value of numeric_only to silence this warning.


_list_series: list[pd.Series] = []
for elem_process in test_lulesh_processes:
    for elem_iteration in test_lulesh_iterations:
        for elem_size in test_lulesh_sizes:
            _list_series_oneCase: list[pd.Series] = []
            for functionName in functionNames_lulesh:
                target_env_oneCase = [
                    (dict_symbols_lulesh["size"], elem_size),
                    (dict_symbols_lulesh["iteration"], elem_iteration),
                    (dict_symbols_lulesh["process"], elem_process),
                ]
                # 全体
                _all: float = (
                    dict_models_lulesh_all[functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )
                # １コール当たり
                _perCall: float = (
                    dict_models_lulesh_perCall[functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )
                # コール回数
                _call: float = (
                    dict_models_lulesh_call[functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )

                target_lulesh_DF: pd.DataFrame = testDF_lulesh[
                    (testDF_lulesh["Name"] == functionName)
                    & (testDF_lulesh["process"] == elem_process)
                    & (testDF_lulesh["iteration"] == elem_iteration)
                    & (testDF_lulesh["size"] == elem_size)
                ]
                target_lulesh_DF = target_lulesh_DF.reset_index()
                # 実測値
                _real_exclusive: float = target_lulesh_DF.loc[0][resVar_all]
                _real_call: float = target_lulesh_DF.loc[0][resVar_call]

                _tmp_series: pd.Series = pd.Series(dtype="object")
                _tmp_series["Name"] = functionName
                _tmp_series["all to all"] = _all
                _tmp_series["perCall to all"] = _perCall * _call
                _tmp_series["real exclusive"] = _real_exclusive
                _tmp_series["real call"] = _real_call

                _list_series_oneCase.append(_tmp_series)

            resultDF: pd.DataFrame = pd.DataFrame(data=_list_series_oneCase)
            resultDF = add_relativeErrorRateCol(
                inputDF=resultDF,
                real_colName="real exclusive",
                predicted_colName="all to all",
                targetColName="vs all to all",
            )
            resultDF = add_relativeErrorRateCol(
                inputDF=resultDF,
                real_colName="real exclusive",
                predicted_colName="perCall to all",
                targetColName="vs perCall to all",
            )
            resultDF["process"] = elem_process
            resultDF["iteration"] = elem_iteration
            resultDF["size"] = elem_size

            resultDF["weighted all"] = calcWeightedMAPEscore(
                inputDF=resultDF,
                inputColumnDict={
                    "funcName": "Name",
                    "call": "real exclusive",
                    "MAPE": "vs all to all",
                },
            )
            resultDF["weighted perCall"] = calcWeightedMAPEscore(
                inputDF=resultDF,
                inputColumnDict={
                    "funcName": "Name",
                    "call": "real exclusive",
                    "MAPE": "vs perCall to all",
                },
            )

            resultDF = resultDF.astype(
                {
                    "all to all": float,
                    "perCall to all": float,
                    "real exclusive": float,
                    "real call": float,
                    "vs all to all": float,
                    "vs perCall to all": float,
                    "weighted perCall": float,
                    "weighted all": float,
                }
            )

            _list_series.append(resultDF.mean(numeric_only=True))

resultDF_ex: pd.DataFrame = pd.DataFrame(_list_series)
resultDF_ex


_list_series: list[pd.Series] = []
for elem_process in train_lulesh_processes:
    for elem_iteration in train_lulesh_iterations:
        for elem_size in train_lulesh_sizes:
            for functionName in functionNames_lulesh:
                target_env_oneCase = [
                    (dict_symbols_lulesh["size"], elem_size),
                    (dict_symbols_lulesh["iteration"], elem_iteration),
                    (dict_symbols_lulesh["process"], elem_process),
                ]

                # 全体
                _all: float = (
                    dict_models_lulesh_all[functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )
                # 1コール当たり
                _perCall: float = (
                    dict_models_lulesh_perCall[functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )
                # コール回数
                _call: float = (
                    dict_models_lulesh_call[functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )

                target_lulesh_DF: pd.DataFrame = trainDF_lulesh[
                    (trainDF_lulesh["Name"] == functionName)
                    & (trainDF_lulesh["process"] == elem_process)
                    & (trainDF_lulesh["iteration"] == elem_iteration)
                    & (trainDF_lulesh["size"] == elem_size)
                ]

                _real_ex: float = target_lulesh_DF.reset_index().loc[0]["Exclusive"]
                _real_call: float = target_lulesh_DF.reset_index().loc[0]["#Call"]

                _tmp_series: pd.Series = pd.Series(
                    {
                        "Name": functionName,
                        "process": elem_process,
                        "iteration": elem_iteration,
                        "size": elem_size,
                        "_all": _all,
                        "_perCall": _perCall * _call,
                        "_real_call": _real_call,
                        "_call": _call,
                        "_real_ex": _real_ex,
                    }
                )

                _list_series.append(_tmp_series)


resultDF_ex_fitness: pd.DataFrame = pd.DataFrame(_list_series)
resultDF_ex_fitness = add_relativeErrorRateCol(
    inputDF=resultDF_ex_fitness,
    real_colName="_real_ex",
    predicted_colName="_all",
    targetColName="vs _all",
)
resultDF_ex_fitness = add_relativeErrorRateCol(
    inputDF=resultDF_ex_fitness,
    real_colName="_real_ex",
    predicted_colName="_perCall",
    targetColName="vs _perCall",
)
resultDF_ex_fitness


resultDF_ex_fitness.mean()

/tmp/ipykernel_121/1762337805.py:1: FutureWarning:

The default value of numeric_only in DataFrame.mean is deprecated. In a future version, it will default to False. In addition, specifying 'numeric_only=None' is deprecated. Select only valid columns or specify the value of numeric_only to silence this warning.

process          130.500000
iteration         49.600000
size              30.000000
_all               0.202899
_perCall           0.186938
_real_call     63915.372350
_call          63915.372350
_real_ex           0.202899
vs _all          152.021226
vs _perCall      106.569120
dtype: float64


print(resultDF_in.to_csv())
print(resultDF_ex.to_csv())

,all to all,perCall to all,real inclusive,real call,vs all to all,vs perCall to all,process,iteration,size,weighted all,weighted perCall
0,10.611351930359527,9.676474217800424,10.839249277552083,2106369.137561035,37.29268669015393,35.43018723269541,512.0,256.0,64.0,7.737989842121091,11.677584614814636
1,35.47642502959257,31.471554773470167,36.179501466729164,7106556.637561034,38.620857939519006,38.064344663643865,512.0,256.0,96.0,12.18557072646931,16.879060132916607
2,83.87770766424927,73.6674641556908,85.97173067903124,16843775.387561034,54.54402041982698,45.268959323177626,512.0,256.0,128.0,15.372071463442103,20.704699677786284
3,22.158231473484413,19.17506225333905,21.59216604190625,4204548.075061034,49.47031161730601,44.051874013721005,512.0,512.0,64.0,12.806894990419677,12.086386051975106
4,74.31459919508686,62.74050156199445,73.69767874593751,14185485.575061036,54.582365471743245,42.422740989603575,512.0,512.0,96.0,18.343713967012658,18.27624453544423
5,175.80302931250344,147.08337889402918,172.9344806705521,33622048.07506103,73.60662845024501,45.30363450313307,512.0,512.0,128.0,20.484699194282967,21.012191216686908
6,48.47662976735646,38.17630899735836,44.16693691421875,8400905.950061036,80.89637897677085,63.477555273759606,512.0,1024.0,64.0,24.250173404713387,14.151127994233732
7,162.6809632707526,125.28247459285637,145.12827248022916,28343280.950061034,89.5643980680369,51.25459148591234,512.0,1024.0,96.0,27.234582765993686,17.369564854515755
8,384.76406064848595,293.91930492411086,344.3190848386354,67178468.45006104,127.67407578608183,53.889614344304164,512.0,1024.0,128.0,30.057984088875145,20.883720467560195
9,10.619840860810292,9.675813824447374,12.003718211395832,2106394.114417867,36.40421419306993,37.32751689271612,729.0,256.0,64.0,15.376220620392743,20.44262556548373
10,35.51258578756824,31.470908313548193,36.52402229557292,7106581.614417867,38.62685545407163,39.60908282616979,729.0,256.0,96.0,12.59448709298471,17.36463000784385
11,83.96775566529726,73.6668448292925,85.82688698654167,16843800.36441787,60.144996553270715,47.124121822515775,729.0,256.0,128.0,15.207500739598695,20.652246235200828
12,22.17837169973441,19.17440186117058,22.966624707375,4204597.968792867,48.67331304789317,39.509350635350664,729.0,512.0,64.0,16.78093243205558,17.16293151612092
13,74.39008307638494,62.73985510325706,74.22244930017709,14185535.468792867,53.314671581624665,41.36237651450493,729.0,512.0,96.0,18.60216198192579,18.647579257837172
14,175.9862876799453,147.08275956881548,172.14807450036457,33622097.96879286,79.30598062379737,48.371852621472414,729.0,512.0,128.0,20.222788208640875,20.8731575283828
15,48.52007258520491,38.17564860578217,44.713030477781246,8401005.681292867,79.89437050531198,59.37217045764684,729.0,1024.0,64.0,25.17095190358091,16.156161021115857
16,162.83509339869548,125.28182813471128,148.1867310167396,28343380.68129287,83.9650752478088,52.108528917455544,729.0,1024.0,96.0,28.213288745997843,18.6858168230563
17,385.13373974871547,293.9186855994894,347.00953282354175,67178568.18129286,118.94803111396571,53.259493741196465,729.0,1024.0,128.0,30.140737908230065,21.168359233315538
18,10.625268769003398,9.675390893871992,11.676582892072915,2106414.43003125,37.71706213556318,37.35015401255731,1000.0,256.0,64.0,13.03755565649311,18.212172234343676
19,35.535707342824665,31.470494292149905,38.129688612802084,7106601.93003125,34.05282840903999,34.52943690409282,1000.0,256.0,96.0,15.27394240214445,20.061783492132943
20,84.02533327009638,73.66644815734435,87.69872022042708,16843820.68003125,51.91755261888863,43.95255639043637,1000.0,256.0,128.0,16.429116481604975,21.834959085378383
21,22.19124956458528,19.173978932074554,23.075291137041667,4204638.55315625,47.84723968883448,39.67141789251325,1000.0,512.0,64.0,17.187381883294037,18.132954438553686
22,74.4383482353613,62.73944108333813,77.02295967070833,14185576.053156251,49.269650020278206,41.0114093774376,1000.0,512.0,96.0,20.790522147394288,20.92657622016308
23,176.10346493800654,147.08236289834664,174.80688687560416,33622138.55315625,73.57784025029191,45.49651504209931,1000.0,512.0,128.0,21.001374894095918,21.694538484423596
24,48.54785036337131,38.17522567742583,46.96365568321875,8401086.79378125,77.62539974192568,60.831822566769915,1000.0,1024.0,64.0,27.50661081606833,19.248749318142803
25,162.93364576511175,125.28141411553202,150.54997028079168,28343461.79378125,80.98162277035146,49.60976001689607,1000.0,1024.0,96.0,28.81203760084329,19.511240434595564
26,385.37011631330097,293.91828892976025,351.66473063015627,67178649.29378125,117.02369439840291,52.677613768154345,1000.0,1024.0,128.0,30.867998998900152,22.073089282079582

,all to all,perCall to all,real exclusive,real call,vs all to all,vs perCall to all,process,iteration,size,weighted all,weighted perCall
0,3.9193650504750863,3.5627250909405053,4.129666003072916,2106369.137561035,33.03513319577337,35.8312710514823,512.0,256.0,64.0,20.74299610609404,19.959975968077426
1,13.114371695548254,11.462402301967256,14.749491093083332,7106556.637561034,43.46161968793186,37.63318659334597,512.0,256.0,96.0,27.631758493085357,27.3352828786892
2,31.020158292806975,26.597047478174648,35.74392863788542,16843775.387561034,61.32428776819555,44.530960698919245,512.0,256.0,128.0,33.29148279558514,33.64580661546467
3,8.713425955234236,7.0863780416345445,8.229332755968752,4204548.075061034,35.54275816866049,56.707941767200815,512.0,512.0,64.0,32.149902642635524,19.9660936835381
4,29.25497442074865,22.87180068370394,29.86740795458334,14185485.575061036,50.78109583579993,42.29548728236925,512.0,512.0,96.0,38.626849250725286,27.92584317326976
5,69.25651630699917,53.11209336731611,71.75545987753125,33622048.07506103,72.43028576527705,59.47950831660272,512.0,512.0,128.0,43.50945201914712,33.60393253550947
6,21.126701856848847,14.139320326223341,16.752916127864584,8400905.950061036,52.64539156573753,107.86259417360145,512.0,1024.0,64.0,55.40986864578554,21.430933614661548
7,71.02739366123484,45.7016446230198,59.08399127064583,28343280.950061034,68.8566544250602,107.17690574090778,512.0,1024.0,96.0,59.68041156080999,27.542671060233797
8,168.2017202741846,106.16151408815065,143.08803279686458,67178468.45006104,102.90053255158602,117.75080155997269,512.0,1024.0,128.0,64.54574492226224,33.72030166712323
9,3.919445980294607,3.5620855607488795,4.470895353062501,2106394.114417867,28.337839200983645,33.31654160280843,729.0,256.0,64.0,19.142831052162933,21.792942027190946
10,13.114452693681553,11.4618143514279,14.86478275171875,7106581.614417867,46.33556420939651,40.23889902711741,729.0,256.0,96.0,27.847072458856,27.522685196478818
11,31.020239423969844,26.59655997222129,35.715866194041666,16843800.36441787,63.040293205839326,43.2292196536778,729.0,256.0,128.0,33.352645253286134,33.77857174767026
12,8.713506885053759,7.085756859042122,8.631676836229166,4204597.968792867,32.98753867140823,46.29985373714845,729.0,512.0,64.0,29.016446155538794,19.8627697401154
13,29.255055418881952,22.8712691002586,30.057959758927087,14185535.468792867,50.99038984567652,42.40075161694834,729.0,512.0,96.0,38.743318693999264,28.098066905932214
14,69.25659743816206,53.11173626642036,71.56764745880209,33622097.96879286,76.38371664113916,53.27036086579582,729.0,512.0,128.0,43.69730071398402,33.82914887910738
15,21.12678278666837,14.138727817658623,16.88281177351042,8401005.681292867,48.95799489003571,101.21595259533208,729.0,1024.0,64.0,50.1517739462712,18.097539337903015
16,71.02747465936814,45.70121775699914,60.05138731569791,28343380.68129287,68.1944256251386,108.78202072822447,729.0,1024.0,96.0,59.83213579517762,28.208945115138505
17,168.20180140534745,106.16140978918948,143.84909536218748,67178568.18129286,99.47995100625835,73.51745301813607,729.0,1024.0,128.0,64.30100707704219,33.69587258655824
18,3.9202046519562193,3.561675970290428,4.377718367802083,2106414.43003125,28.275865659897796,33.86424685084835,1000.0,256.0,64.0,17.709084435526226,20.27572422311563
19,13.115211409023686,11.461437741522039,15.337157404572915,7106601.93003125,40.035304403919326,38.153890169480235,1000.0,256.0,96.0,28.767565685681113,28.708201903182744
20,31.020998224372434,26.596247587602043,36.279199431052085,16843820.68003125,61.21286496045806,43.46128136542685,1000.0,256.0,128.0,33.72820602404703,34.07014799307199
21,8.714265556715368,7.08535900094654,8.668155767145834,4204638.55315625,36.3640714227881,47.25396923396612,1000.0,512.0,64.0,32.80962789266332,19.992637932307467
22,29.25581413422408,22.87092853276604,30.877772213312497,14185576.053156251,49.03844910558851,45.953219636374385,1000.0,512.0,96.0,39.50625321798285,29.16820670962302
23,69.25735623856464,53.111507264838956,72.3799390005,33622138.55315625,71.82952633252792,48.82069879725345,1000.0,512.0,128.0,43.82367513571072,34.00641063896023
24,21.12754145832998,14.138348294388027,17.585509895614585,8401086.79378125,45.92734540628097,100.9387766416295,1000.0,1024.0,64.0,50.567380831185126,21.253764969989863
25,71.02823337471028,45.70094414725054,60.735022405062494,28343461.79378125,66.11471273602014,105.33127016716732,1000.0,1024.0,96.0,59.71322678328758,28.463254338695744
26,168.20256020575005,106.16134243208899,145.37338692223958,67178649.29378125,98.42000324945845,76.18677336172527,1000.0,1024.0,128.0,64.50711017592604,34.16517700761053


# print(resultDF_in_fitness.dtypes)
# resultDF_in_fitness = resultDF_in_fitness.astype(
#     {"vs _all": float, "vs _perCall": float, "_call": float}
# )
# fig = px.scatter(resultDF_in_fitness, x="_call", y="vs _perCall")
# fig.show()


resultDF_ex_fitness = resultDF_ex_fitness.astype(
    {"vs _all": float, "vs _perCall": float, "_call": float}
)
fig = px.scatter(resultDF_ex_fitness, x="_call", y="vs _perCall")
fig.show()


# exclusive, モデル適合度

_list_series: list[pd.Series] = []
for functionName in functionNames_lulesh:
    _list_series_oneCase: list[pd.Series] = []
    for elem_process in train_lulesh_processes:
        for elem_iteration in train_lulesh_iterations:
            for elem_size in train_lulesh_sizes:
                target_env_oneCase = [
                    (dict_symbols_lulesh["size"], elem_size),
                    (dict_symbols_lulesh["iteration"], elem_iteration),
                    (dict_symbols_lulesh["process"], elem_process),
                ]
                # 全体
                _all: float = (
                    dict_models_lulesh_all[functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )
                # １コール当たり
                _perCall: float = (
                    dict_models_lulesh_perCall[functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )
                # コール回数
                _call: float = (
                    dict_models_lulesh_call[functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )

                target_lulesh_DF: pd.DataFrame = trainDF_lulesh[
                    (trainDF_lulesh["Name"] == functionName)
                    & (trainDF_lulesh["process"] == elem_process)
                    & (trainDF_lulesh["iteration"] == elem_iteration)
                    & (trainDF_lulesh["size"] == elem_size)
                ]
                target_lulesh_DF = target_lulesh_DF.reset_index()
                # 実測値
                _real_exclusive: float = target_lulesh_DF.loc[0][resVar_all]
                _real_call: float = target_lulesh_DF.loc[0][resVar_call]

                _tmp_series: pd.Series = pd.Series(dtype="object")
                _tmp_series["Name"] = functionName
                _tmp_series["all to all"] = _all
                _tmp_series["perCall to all"] = _perCall * _call
                _tmp_series["real exclusive"] = _real_exclusive
                _tmp_series["real call"] = _real_call

                _list_series_oneCase.append(_tmp_series)

    resultDF: pd.DataFrame = pd.DataFrame(data=_list_series_oneCase)
    resultDF = add_relativeErrorRateCol(
        inputDF=resultDF,
        real_colName="real exclusive",
        predicted_colName="all to all",
        targetColName="vs all to all",
    )
    resultDF = add_relativeErrorRateCol(
        inputDF=resultDF,
        real_colName="real exclusive",
        predicted_colName="perCall to all",
        targetColName="vs perCall to all",
    )

    resultDF = resultDF.astype(
        {
            "all to all": float,
            "perCall to all": float,
            "real exclusive": float,
            "real call": float,
            "vs all to all": float,
            "vs perCall to all": float,
        }
    )

    _tmp_list_series: pd.Series = resultDF.mean(numeric_only=True)
    _tmp_list_series["Name"] = functionName
    _list_series.append(_tmp_list_series)

resultDF_ex_fitness: pd.DataFrame = pd.DataFrame(_list_series)
resultDF_ex_fitness


print(resultDF_ex_fitness.to_csv())

,all to all,perCall to all,real exclusive,real call,vs all to all,vs perCall to all,Name
0,3.5870833333333345e-06,3.6112499999999944e-06,3.5870833333333332e-06,1.0,4.49497172030504,4.529863722075408,.TAU_application
1,0.4656926444455586,0.49987631194890925,0.4656926444444444,48.6,64.18695938041436,61.08904659123899,MPI_Allreduce()
2,0.0008163305555555552,0.0008869008333333322,0.0008163305555555556,1.0,375.84537038593464,413.75958373384805,MPI_Barrier()
3,0.00019597249999999986,0.00020896345164618952,0.00019597250000000002,451.4,13.913296492896276,21.117242604652933,MPI_Comm_rank()
4,4.641972222222224e-07,4.5992499999999945e-07,4.6419722222222224e-07,1.0,14.603986505879087,14.554590269120544,MPI_Comm_size()
5,2.3415972222222217,1.8649083333325853,2.341597222222222,1.0,58.9390413018782,76.74628114262062,MPI_Finalize()
6,0.4930194444444449,0.4953749999999999,0.49301944444444445,1.0,22.84095182616433,22.984320327785227,MPI_Init()
7,0.0008281572222222218,0.000888816204379018,0.0008281572222222223,1310.1943333333334,29.90513035960323,34.62477168068254,MPI_Irecv()
8,0.008510761111111062,0.008991201163332822,0.00851076111111111,1310.1943333333334,194.1246496851199,173.78932859192264,MPI_Isend()
9,0.001277368055555555,0.0011958291666666653,0.0012773680555555555,1.0,680.7164258821371,609.2832085819716,MPI_Reduce()
10,0.07208439194444444,0.07559656193233581,0.07208439194444445,1310.1943333333334,637.5369229766028,676.3712674322157,MPI_Wait()
11,0.23594427222231978,0.2651486611378626,0.23594427222222222,149.8,238.4127177258358,229.11891904337725,MPI_Waitall()
12,0.31019236388881327,0.3055593054414882,0.3101923638888889,2040193.5,16.134982581537752,15.74330090561876,Real_t_CalcElemVolume(const
13,1.918166666666667e-06,1.8962499999999975e-06,1.9181666666666667e-06,2.0,4.813320856564633,4.558067546197548,StrToInt
14,1.9968472222222227,1.9846333333333317,1.9968472222222222,1.0,7.036761534230096,13.104779042937851,int_main(int_char_**)
15,0.44255354722222234,0.4260045833333333,0.44255354722222223,49.6,63.54839751213876,16.200636428200706,void_CalcKinematicsForElems(Domain
16,0.005124422777779533,0.0033976450985192734,0.005124422777777778,49.6,55.767147199354824,37.61482753419218,void_CommMonoQ(Domain
17,0.000546600833333333,0.0003667987687112887,0.0005466008333333332,149.8,189.18097607176196,22.339178335592205,void_CommRecv(Domain
18,0.009188817222222223,0.0038924756795904623,0.009188817222222223,50.6,411.3825610915216,39.75305364159599,void_CommSBN(Domain
19,0.07149106388888887,0.016184991152118418,0.07149106388888889,149.8,847.5482819714033,54.300344562875615,void_CommSend(Domain
20,0.021035447222222222,0.012811309257011775,0.021035447222222222,49.6,147.48972108475215,83.518617100871,void_CommSyncPosVel(Domain
21,0.0005171552777777776,0.00047565666666666575,0.0005171552777777778,1.0,25.108522548994234,29.898369026127284,void_Domain::BuildMesh(Int_t_Int_t_Int_t)
22,0.00040439499999994334,0.00041608333333320954,0.0004043950000000001,1.0,9.228405744946002,9.643183299008875,void_Domain::CreateRegionIndexSets(Int_t_Int_t)
23,0.014089474999999995,0.014358333333333221,0.014089475,1.0,29.694413955518637,23.95741877715,void_Domain::Domain(Int_t_Index_t_Index_t_Index_t_Index_t_Int_t_Int_t_Int_t_Int_t)
24,8.139061111461348e-05,7.854016667015599e-05,8.139061111111112e-05,1.0,61.17071685512961,56.710330624974965,void_Domain::SetupBoundaryConditions(Int_t)
25,0.00034043472222222216,0.0003449474999999997,0.00034043472222222216,1.0,25.141754505135584,25.130508415523774,void_Domain::SetupCommBuffers(Int_t)
26,0.0003176191666666666,0.0003211899999999997,0.0003176191666666667,1.0,4.97185554826168,5.742695863289117,void_Domain::SetupElementConnectivities(Int_t)
27,2.7751999999999993e-06,2.7767499999999963e-06,2.7752e-06,1.0,6.611437460710096,6.463982656264216,void_Domain::SetupSymmetryPlanes(Int_t)
28,5.8533444444444365e-05,5.811649999999982e-05,5.853344444444444e-05,1.0,587.8876152326413,591.6323906857891,void_Domain::~Domain()
29,1.1277055555555564e-05,1.1272333333333327e-05,1.1277055555555559e-05,1.0,5.037709611570846,5.035600090680422,void_InitMeshDecomp(Int_t_Int_t_Int_t
30,2.085277777777778e-06,2.0605833333333323e-06,2.085277777777778e-06,1.0,25.389755169599674,24.495431404947393,void_ParseCommandLineOptions(int_char_**_Int_t_cmdLineOpts
31,2.7425583333333326e-06,2.827912142614843e-06,2.7425583333333335e-06,0.03220118000000001,6.014462193451478,6.400693054090221,void_VerifyAndWriteFinalOutput(Real_t_Domain


resultDF_ex_fitness.mean()

/tmp/ipykernel_121/3761618132.py:1: FutureWarning:

The default value of numeric_only in DataFrame.mean is deprecated. In a future version, it will default to False. In addition, specifying 'numeric_only=None' is deprecated. Select only valid columns or specify the value of numeric_only to silence this warning.

all to all               0.202899
perCall to all           0.186938
real exclusive           0.202899
real call            63915.372350
vs all to all          152.021226
vs perCall to all      106.569120
dtype: float64


print(resultDF_in_fitness.to_csv())

,all to all,perCall to all,real inclusive,real call,vs all to all,vs perCall to all,Name
0,6.499274999999998,5.99680000000455,6.499275,1.0,18.531207791142045,47.868233037072244,.TAU_application
1,0.4851583333338704,0.49987631194890925,0.4851583333333334,48.6,70.02850227605273,54.30365103330267,MPI_Allreduce()
2,0.0008691769444444439,0.0008869008333333322,0.0008691769444444445,1.0,396.74788673794706,409.61022416914784,MPI_Barrier()
3,0.00019597805555555544,0.00020896345164618952,0.00019597805555555557,451.4,13.912199957407852,21.113673856714986,MPI_Comm_rank()
4,4.641972222222224e-07,4.5992499999999945e-07,4.6419722222222224e-07,1.0,14.603986505879087,14.554590269120544,MPI_Comm_size()
5,2.350408333333333,1.8649083333325853,2.350408333333333,1.0,57.047656270041564,72.38078156462822,MPI_Finalize()
6,0.5034055555555559,0.4953749999999999,0.5034055555555556,1.0,19.225513434065896,18.808968818305168,MPI_Init()
7,0.0008598322222222225,0.000888816204379018,0.0008598322222222223,1310.1943333333334,30.815302940609747,33.88940863615063,MPI_Irecv()
8,0.010041902777777827,0.008991201163332822,0.010041902777777778,1310.1943333333334,196.52093502405054,168.8005687628724,MPI_Isend()
9,0.0013403763888888887,0.0011958291666666653,0.001340376388888889,1.0,711.2399128509621,608.3507686480746,MPI_Reduce()
10,0.0803010583333331,0.07559656193233581,0.08030105833333333,1310.1943333333334,670.2231152824589,593.3711312685786,MPI_Wait()
11,0.25203038333343813,0.2651486611378626,0.25203038333333333,149.8,235.4690990927873,227.9093300733319,MPI_Waitall()
12,0.3230555555555556,0.3055593054414882,0.32305555555555554,2040193.5,26.6534778405902,9.741669563749706,Real_t_CalcElemVolume(const
13,1.918166666666667e-06,1.8962499999999975e-06,1.9181666666666667e-06,2.0,4.813320856564633,4.558067546197548,StrToInt
14,6.4992722222222215,5.99680000000455,6.4992722222222215,1.0,18.531234658512858,47.86843209407313,int_main(int_char_**)
15,0.7321944444444437,0.7243828124999999,0.7321944444444444,49.6,20.14709621116764,8.70147828104974,void_CalcKinematicsForElems(Domain
16,0.019962760000000277,0.021199564416666674,0.019962760000000003,49.6,337.411413990948,587.2208886777314,void_CommMonoQ(Domain
17,0.0013260472222222223,0.0012621483763984467,0.0013260472222222225,149.8,28.323249555554177,27.54347875087964,void_CommRecv(Domain
18,0.032381750000000015,0.035455116610314086,0.03238175,50.6,653.7729035533267,647.3330877572415,void_CommSBN(Domain
19,0.29726666666666707,0.29954774215774066,0.2972666666666666,149.8,175.67952170188036,174.1534561960157,void_CommSend(Domain
20,0.051118394444444215,0.04868275385729218,0.05111839444444444,49.6,187.35303190142466,142.852550568504,void_CommSyncPosVel(Domain
21,0.0005212413888888888,0.00047565666666666575,0.000521241388888889,1.0,26.816747878454173,30.11944842766424,void_Domain::BuildMesh(Int_t_Int_t_Int_t)
22,0.0004200038888887566,0.0004166691666666678,0.000420003888888889,1.0,7.963391919754539,7.35237109953441,void_Domain::CreateRegionIndexSets(Int_t_Int_t)
23,0.02445429722222223,0.023516666666666592,0.02445429722222222,1.0,16.558679738181976,12.883915270762701,void_Domain::Domain(Int_t_Index_t_Index_t_Index_t_Index_t_Int_t_Int_t_Int_t_Int_t)
24,8.139061111461348e-05,7.854016667015599e-05,8.139061111111112e-05,1.0,61.17071685512961,56.710330624974965,void_Domain::SetupBoundaryConditions(Int_t)
25,0.00034736777777777774,0.0003449474999999997,0.00034736777777777774,1.0,22.790014098984237,22.831558767389716,void_Domain::SetupCommBuffers(Int_t)
26,0.00032199138888888896,0.0003211899999999997,0.0003219913888888889,1.0,5.643632463244711,5.133162906038845,void_Domain::SetupElementConnectivities(Int_t)
27,2.786811111111111e-06,2.7767499999999963e-06,2.7868111111111116e-06,1.0,6.538503805788273,6.346428561026772,void_Domain::SetupSymmetryPlanes(Int_t)
28,5.8533444444444365e-05,5.811649999999982e-05,5.853344444444444e-05,1.0,587.8876152326413,591.6323906857891,void_Domain::~Domain()
29,1.1277055555555564e-05,1.1272333333333327e-05,1.1277055555555559e-05,1.0,5.037709611570846,5.035600090680422,void_InitMeshDecomp(Int_t_Int_t_Int_t
30,3.947027777777779e-06,3.957999999999994e-06,3.947027777777778e-06,1.0,12.453268707970874,12.553922598018122,void_ParseCommandLineOptions(int_char_**_Int_t_cmdLineOpts
31,2.787308333333331e-06,2.827912142614843e-06,2.7873083333333333e-06,0.03220118000000001,9.054462425440855,5.655438325067181,void_VerifyAndWriteFinalOutput(Real_t_Domain


resultDF_in_fitness.mean()

/tmp/ipykernel_121/2420836859.py:1: FutureWarning:

The default value of numeric_only in DataFrame.mean is deprecated. In a future version, it will default to False. In addition, specifying 'numeric_only=None' is deprecated. Select only valid columns or specify the value of numeric_only to silence this warning.

all to all               0.567709
perCall to all           0.520875
real inclusive           0.567709
real call            63915.372350
vs all to all          145.280166
vs perCall to all      146.162156
dtype: float64


# fig_ex_perCall = px.scatter(
#     resultDF_ex_fitness,
#     x="Name",
#     y="vs _perCall",
#     # size="_call",
#     labels={
#         "Name": "関数名",
#         "vs _perCall": "MAPE",
#     },
#     hover_data=["_real_call"],
# )
# fig_ex_perCall.show()

# fig_ex_all = px.scatter(
#     resultDF_ex_fitness,
#     x="Name",
#     y="vs _all",
#     # size="_call",
#     labels={
#         "Name": "関数名",
#         "vs _perCall": "MAPE",
#     },
#     hover_data=["_real_call"],
# )
# fig_ex_all.show()


# fig_in_perCall = px.scatter(
#     resultDF_in_fitness,
#     x="Name",
#     y="vs _perCall",
#     labels={
#         "Name": "関数名",
#         "vs _perCall": "MAPE",
#     },
#     hover_data=["_real_call"],
# )
# fig_in_perCall.show()

# fig_in_all = px.scatter(
#     resultDF_in_fitness,
#     x="Name",
#     y="vs _all",
#     labels={
#         "Name": "関数名",
#         "vs _perCall": "MAPE",
#     },
#     hover_data=["_real_call"],
# )
# fig_in_all.show()


print(resultDF_ex.mean())

print(resultDF_in.mean())

all to all           4.618193e+01
perCall to all       3.229901e+01
real exclusive       4.300393e+01
real call            2.022133e+07
vs all to all        5.677421e+01
vs perCall to all    6.264829e+01
process              7.470000e+02
iteration            5.973333e+02
size                 9.600000e+01
weighted all         4.121500e+01
weighted perCall     2.741189e+01
dtype: float64
all to all           1.110029e+02
perCall to all       8.902083e+01
real inclusive       1.051859e+02
real call            2.022133e+07
vs all to all        6.576082e+01
vs perCall to all    4.601254e+01
process              7.470000e+02
iteration            5.973333e+02
size                 9.600000e+01
weighted all         2.006257e+01
weighted perCall     1.873667e+01
dtype: float64


print(resultDF_ex.to_csv())
resultDF_ex

,all to all,perCall to all,real exclusive,real call,vs all to all,vs perCall to all,process,iteration,size,weighted all,weighted perCall
0,3.9193650504750863,3.5627250909405053,4.129666003072916,2106369.137561035,33.03513319577337,35.8312710514823,512.0,256.0,64.0,20.74299610609404,19.959975968077426
1,13.114371695548254,11.462402301967256,14.749491093083332,7106556.637561034,43.46161968793186,37.63318659334597,512.0,256.0,96.0,27.631758493085357,27.3352828786892
2,31.020158292806975,26.597047478174648,35.74392863788542,16843775.387561034,61.32428776819555,44.530960698919245,512.0,256.0,128.0,33.29148279558514,33.64580661546467
3,8.713425955234236,7.0863780416345445,8.229332755968752,4204548.075061034,35.54275816866049,56.707941767200815,512.0,512.0,64.0,32.149902642635524,19.9660936835381
4,29.25497442074865,22.87180068370394,29.86740795458334,14185485.575061036,50.78109583579993,42.29548728236925,512.0,512.0,96.0,38.626849250725286,27.92584317326976
5,69.25651630699917,53.11209336731611,71.75545987753125,33622048.07506103,72.43028576527705,59.47950831660272,512.0,512.0,128.0,43.50945201914712,33.60393253550947
6,21.126701856848847,14.139320326223341,16.752916127864584,8400905.950061036,52.64539156573753,107.86259417360145,512.0,1024.0,64.0,55.40986864578554,21.430933614661548
7,71.02739366123484,45.7016446230198,59.08399127064583,28343280.950061034,68.8566544250602,107.17690574090778,512.0,1024.0,96.0,59.68041156080999,27.542671060233797
8,168.2017202741846,106.16151408815065,143.08803279686458,67178468.45006104,102.90053255158602,117.75080155997269,512.0,1024.0,128.0,64.54574492226224,33.72030166712323
9,3.919445980294607,3.5620855607488795,4.470895353062501,2106394.114417867,28.337839200983645,33.31654160280843,729.0,256.0,64.0,19.142831052162933,21.792942027190946
10,13.114452693681553,11.4618143514279,14.86478275171875,7106581.614417867,46.33556420939651,40.23889902711741,729.0,256.0,96.0,27.847072458856,27.522685196478818
11,31.020239423969844,26.59655997222129,35.715866194041666,16843800.36441787,63.040293205839326,43.2292196536778,729.0,256.0,128.0,33.352645253286134,33.77857174767026
12,8.713506885053759,7.085756859042122,8.631676836229166,4204597.968792867,32.98753867140823,46.29985373714845,729.0,512.0,64.0,29.016446155538794,19.8627697401154
13,29.255055418881952,22.8712691002586,30.057959758927087,14185535.468792867,50.99038984567652,42.40075161694834,729.0,512.0,96.0,38.743318693999264,28.098066905932214
14,69.25659743816206,53.11173626642036,71.56764745880209,33622097.96879286,76.38371664113916,53.27036086579582,729.0,512.0,128.0,43.69730071398402,33.82914887910738
15,21.12678278666837,14.138727817658623,16.88281177351042,8401005.681292867,48.95799489003571,101.21595259533208,729.0,1024.0,64.0,50.1517739462712,18.097539337903015
16,71.02747465936814,45.70121775699914,60.05138731569791,28343380.68129287,68.1944256251386,108.78202072822447,729.0,1024.0,96.0,59.83213579517762,28.208945115138505
17,168.20180140534745,106.16140978918948,143.84909536218748,67178568.18129286,99.47995100625835,73.51745301813607,729.0,1024.0,128.0,64.30100707704219,33.69587258655824
18,3.9202046519562193,3.561675970290428,4.377718367802083,2106414.43003125,28.275865659897796,33.86424685084835,1000.0,256.0,64.0,17.709084435526226,20.27572422311563
19,13.115211409023686,11.461437741522039,15.337157404572915,7106601.93003125,40.035304403919326,38.153890169480235,1000.0,256.0,96.0,28.767565685681113,28.708201903182744
20,31.020998224372434,26.596247587602043,36.279199431052085,16843820.68003125,61.21286496045806,43.46128136542685,1000.0,256.0,128.0,33.72820602404703,34.07014799307199
21,8.714265556715368,7.08535900094654,8.668155767145834,4204638.55315625,36.3640714227881,47.25396923396612,1000.0,512.0,64.0,32.80962789266332,19.992637932307467
22,29.25581413422408,22.87092853276604,30.877772213312497,14185576.053156251,49.03844910558851,45.953219636374385,1000.0,512.0,96.0,39.50625321798285,29.16820670962302
23,69.25735623856464,53.111507264838956,72.3799390005,33622138.55315625,71.82952633252792,48.82069879725345,1000.0,512.0,128.0,43.82367513571072,34.00641063896023
24,21.12754145832998,14.138348294388027,17.585509895614585,8401086.79378125,45.92734540628097,100.9387766416295,1000.0,1024.0,64.0,50.567380831185126,21.253764969989863
25,71.02823337471028,45.70094414725054,60.735022405062494,28343461.79378125,66.11471273602014,105.33127016716732,1000.0,1024.0,96.0,59.71322678328758,28.463254338695744
26,168.20256020575005,106.16134243208899,145.37338692223958,67178649.29378125,98.42000324945845,76.18677336172527,1000.0,1024.0,128.0,64.50711017592604,34.16517700761053


print(resultDF_in.to_csv())
resultDF_in

,all to all,perCall to all,real inclusive,real call,vs all to all,vs perCall to all,process,iteration,size,weighted all,weighted perCall
0,10.611351930359527,9.676474217800424,10.839249277552083,2106369.137561035,37.29268669015393,35.43018723269541,512.0,256.0,64.0,7.737989842121091,11.677584614814636
1,35.47642502959257,31.471554773470167,36.179501466729164,7106556.637561034,38.620857939519006,38.064344663643865,512.0,256.0,96.0,12.18557072646931,16.879060132916607
2,83.87770766424927,73.6674641556908,85.97173067903124,16843775.387561034,54.54402041982698,45.268959323177626,512.0,256.0,128.0,15.372071463442103,20.704699677786284
3,22.158231473484413,19.17506225333905,21.59216604190625,4204548.075061034,49.47031161730601,44.051874013721005,512.0,512.0,64.0,12.806894990419677,12.086386051975106
4,74.31459919508686,62.74050156199445,73.69767874593751,14185485.575061036,54.582365471743245,42.422740989603575,512.0,512.0,96.0,18.343713967012658,18.27624453544423
5,175.80302931250344,147.08337889402918,172.9344806705521,33622048.07506103,73.60662845024501,45.30363450313307,512.0,512.0,128.0,20.484699194282967,21.012191216686908
6,48.47662976735646,38.17630899735836,44.16693691421875,8400905.950061036,80.89637897677085,63.477555273759606,512.0,1024.0,64.0,24.250173404713387,14.151127994233732
7,162.6809632707526,125.28247459285637,145.12827248022916,28343280.950061034,89.5643980680369,51.25459148591234,512.0,1024.0,96.0,27.234582765993686,17.369564854515755
8,384.76406064848595,293.91930492411086,344.3190848386354,67178468.45006104,127.67407578608183,53.889614344304164,512.0,1024.0,128.0,30.057984088875145,20.883720467560195
9,10.619840860810292,9.675813824447374,12.003718211395832,2106394.114417867,36.40421419306993,37.32751689271612,729.0,256.0,64.0,15.376220620392743,20.44262556548373
10,35.51258578756824,31.470908313548193,36.52402229557292,7106581.614417867,38.62685545407163,39.60908282616979,729.0,256.0,96.0,12.59448709298471,17.36463000784385
11,83.96775566529726,73.6668448292925,85.82688698654167,16843800.36441787,60.144996553270715,47.124121822515775,729.0,256.0,128.0,15.207500739598695,20.652246235200828
12,22.17837169973441,19.17440186117058,22.966624707375,4204597.968792867,48.67331304789317,39.509350635350664,729.0,512.0,64.0,16.78093243205558,17.16293151612092
13,74.39008307638494,62.73985510325706,74.22244930017709,14185535.468792867,53.314671581624665,41.36237651450493,729.0,512.0,96.0,18.60216198192579,18.647579257837172
14,175.9862876799453,147.08275956881548,172.14807450036457,33622097.96879286,79.30598062379737,48.371852621472414,729.0,512.0,128.0,20.222788208640875,20.8731575283828
15,48.52007258520491,38.17564860578217,44.713030477781246,8401005.681292867,79.89437050531198,59.37217045764684,729.0,1024.0,64.0,25.17095190358091,16.156161021115857
16,162.83509339869548,125.28182813471128,148.1867310167396,28343380.68129287,83.9650752478088,52.108528917455544,729.0,1024.0,96.0,28.213288745997843,18.6858168230563
17,385.13373974871547,293.9186855994894,347.00953282354175,67178568.18129286,118.94803111396571,53.259493741196465,729.0,1024.0,128.0,30.140737908230065,21.168359233315538
18,10.625268769003398,9.675390893871992,11.676582892072915,2106414.43003125,37.71706213556318,37.35015401255731,1000.0,256.0,64.0,13.03755565649311,18.212172234343676
19,35.535707342824665,31.470494292149905,38.129688612802084,7106601.93003125,34.05282840903999,34.52943690409282,1000.0,256.0,96.0,15.27394240214445,20.061783492132943
20,84.02533327009638,73.66644815734435,87.69872022042708,16843820.68003125,51.91755261888863,43.95255639043637,1000.0,256.0,128.0,16.429116481604975,21.834959085378383
21,22.19124956458528,19.173978932074554,23.075291137041667,4204638.55315625,47.84723968883448,39.67141789251325,1000.0,512.0,64.0,17.187381883294037,18.132954438553686
22,74.4383482353613,62.73944108333813,77.02295967070833,14185576.053156251,49.269650020278206,41.0114093774376,1000.0,512.0,96.0,20.790522147394288,20.92657622016308
23,176.10346493800654,147.08236289834664,174.80688687560416,33622138.55315625,73.57784025029191,45.49651504209931,1000.0,512.0,128.0,21.001374894095918,21.694538484423596
24,48.54785036337131,38.17522567742583,46.96365568321875,8401086.79378125,77.62539974192568,60.831822566769915,1000.0,1024.0,64.0,27.50661081606833,19.248749318142803
25,162.93364576511175,125.28141411553202,150.54997028079168,28343461.79378125,80.98162277035146,49.60976001689607,1000.0,1024.0,96.0,28.81203760084329,19.511240434595564
26,385.37011631330097,293.91828892976025,351.66473063015627,67178649.29378125,117.02369439840291,52.677613768154345,1000.0,1024.0,128.0,30.867998998900152,22.073089282079582


dict_functionName_model_call

{'.TAU_application': 0.9999999999999989,
 'MPI_Allreduce()': 69114.41388888886,
 'MPI_Barrier()': 0.6301048971915498*process + 32.380652583169415,
 'MPI_Comm_rank()': 7.478305319873806*iteration + 24.533347800925743,
 'MPI_Comm_size()': 0.9889531893055545,
 'MPI_Finalize()': 32774.5344861112,
 'MPI_Init()': 30426.886305555618,
 'MPI_Irecv()': -0.003829264711374236*iteration*size**2 + 26.240016017329342*iteration - 38.59922858812197,
 'MPI_Isend()': 25.586286063268066*iteration - 108.84110916015148*iteration/process - 61.42217708333335,
 'MPI_Reduce()': 0.016384459516614158*iteration*process - 0.35540347438174624*process + 34.33414884767836,
 'MPI_Wait()': 5079.973744444446,
 'MPI_Waitall()': 0.9376670852678016*size**3 + 9342.44998311334,
 'Real_t_CalcElemVolume(const': 0.9478893919798221*iteration*size**3 - 2.421500062860172*size**3 - 29758.051380170076,
 'StrToInt': 0.0036904761904761624*size + 1.8253968253968273,
 'int_main(int_char_**)': 0.9999999999999989,
 'void_CalcKinematicsForElems(Domain': 1526.1962444444437,
 'void_CommMonoQ(Domain': 0.00851523867407278*iteration*process + 89.41085477712845,
 'void_CommRecv(Domain': 3.432252907519414*iteration - 4.979994212962981,
 'void_CommSBN(Domain': 366.1933833333328,
 'void_CommSend(Domain': 81149.40562222223,
 'void_CommSyncPosVel(Domain': 6030.22978611111,
 'void_Domain::BuildMesh(Int_t_Int_t_Int_t)': 43.86055833333329,
 'void_Domain::CreateRegionIndexSets(Int_t_Int_t)': 7.608888888888915,
 'void_Domain::Domain(Int_t_Index_t_Index_t_Index_t_Index_t_Int_t_Int_t_Int_t_Int_t)': 293.55321944444415,
 'void_Domain::SetupBoundaryConditions(Int_t)': 5.056377777777794,
 'void_Domain::SetupCommBuffers(Int_t)': 19.522633333333292,
 'void_Domain::SetupElementConnectivities(Int_t)': 1.1540446825396833*size - 16.749323809523876,
 'void_Domain::SetupSymmetryPlanes(Int_t)': 1.0111111111111102,
 'void_Domain::~Domain()': 1.8583767361111097,
 'void_InitMeshDecomp(Int_t_Int_t_Int_t': 0.9607767489722213,
 'void_ParseCommandLineOptions(int_char_**_Int_t_cmdLineOpts': 0.9786008222222212,
 'void_VerifyAndWriteFinalOutput(Real_t_Domain': 0.10938257227777808}


_list_series: list[pd.Series] = []
for elem_process in train_lulesh_processes:
    for elem_iteration in train_lulesh_iterations:
        for elem_size in train_lulesh_sizes:
            for functionName in functionNames_lulesh:
                target_env_oneCase = [
                    (dict_symbols_lulesh["size"], elem_size),
                    (dict_symbols_lulesh["iteration"], elem_iteration),
                    (dict_symbols_lulesh["process"], elem_process),
                ]

                # コール回数
                _predicted_call: float = (
                    dict_models_lulesh_call[functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )

                target_lulesh_DF: pd.DataFrame = trainDF_lulesh[
                    (trainDF_lulesh["Name"] == functionName)
                    & (trainDF_lulesh["process"] == elem_process)
                    & (trainDF_lulesh["iteration"] == elem_iteration)
                    & (trainDF_lulesh["size"] == elem_size)
                ]
                _real_call: float = target_lulesh_DF.reset_index().loc[0]["#Call"]

                _tmp_series: pd.Series = pd.Series(
                    {
                        "Name": functionName,
                        "process": elem_process,
                        "iteration": elem_iteration,
                        "size": elem_size,
                        "_predicted_call": _predicted_call,
                        "_real_call": _real_call,
                    }
                )

                _list_series.append(_tmp_series)

resultDF_call_fitness: pd.DataFrame = pd.DataFrame(data=_list_series)

resultDF_call_fitness = add_relativeErrorRateCol(
    inputDF=resultDF_call_fitness,
    real_colName="_real_call",
    predicted_colName="_predicted_call",
    targetColName="call relative error rate",
)
resultDF_call_fitness


_list_series: list[pd.Series] = []
for functionName in functionNames_lulesh:
    resultDF_call_fitness_perFunc: pd.DataFrame = resultDF_call_fitness[
        resultDF_call_fitness["Name"] == functionName
    ]
    resultDF_call_fitness_perFunc = resultDF_call_fitness_perFunc.astype(
        {
            "process": float,
            "iteration": float,
            "size": float,
            "_predicted_call": float,
            "_real_call": float,
            "call relative error rate": float,
        }
    )
    _tmp_series: pd.Series = resultDF_call_fitness_perFunc.mean(numeric_only=True)
    _tmp_series["Name"] = functionName
    _list_series.append(_tmp_series)
pd.options.display.float_format = "{:.4g}".format
_tmp_DF: pd.DataFrame = (
    pd.DataFrame(data=_list_series)
    .drop(columns=["process", "iteration", "size"])
    .rename(
        columns={
            "_predicted_call": "予測値",
            "_real_call": "実測値",
            "call relative error rate": "MAPE",
            "Name": "関数名",
        }
    )
)
_tmp_DF


print(_tmp_DF.to_csv())

,予測値,実測値,MAPE,関数名
0,0.9999999999999987,1.0,1.1102230246251565e-13,.TAU_application
1,48.599999999999994,48.6,4.0563281135330706e-14,MPI_Allreduce()
2,0.9999999999999987,1.0,1.1102230246251565e-13,MPI_Barrier()
3,451.40000000000015,451.4,6.454586758761107e-14,MPI_Comm_rank()
4,0.9999999999999987,1.0,1.1102230246251565e-13,MPI_Comm_size()
5,0.9999999999999987,1.0,1.1102230246251565e-13,MPI_Finalize()
6,0.9999999999999987,1.0,1.1102230246251565e-13,MPI_Init()
7,1310.1943333333334,1310.1943333333334,28.141797322952478,MPI_Irecv()
8,1310.1943333333334,1310.1943333333334,28.141797322952478,MPI_Isend()
9,0.9999999999999987,1.0,1.1102230246251565e-13,MPI_Reduce()
10,1310.1943333333334,1310.1943333333334,28.141797322952478,MPI_Wait()
11,149.79999999999993,149.8,3.219714529811956e-14,MPI_Waitall()
12,2040193.4999999986,2040193.5,0.0004922552347244996,Real_t_CalcElemVolume(const
13,1.9999999999999973,2.0,1.1102230246251565e-13,StrToInt
14,0.9999999999999987,1.0,1.1102230246251565e-13,int_main(int_char_**)
15,49.6,49.6,0.0,void_CalcKinematicsForElems(Domain
16,49.6,49.6,0.0,void_CommMonoQ(Domain
17,149.79999999999993,149.8,3.219714529811956e-14,void_CommRecv(Domain
18,50.599999999999994,50.6,1.703281651373677e-14,void_CommSBN(Domain
19,149.79999999999993,149.8,3.219714529811956e-14,void_CommSend(Domain
20,49.6,49.6,0.0,void_CommSyncPosVel(Domain
21,0.9999999999999987,1.0,1.1102230246251565e-13,void_Domain::BuildMesh(Int_t_Int_t_Int_t)
22,0.9999999999999987,1.0,1.1102230246251565e-13,void_Domain::CreateRegionIndexSets(Int_t_Int_t)
23,0.9999999999999987,1.0,1.1102230246251565e-13,void_Domain::Domain(Int_t_Index_t_Index_t_Index_t_Index_t_Int_t_Int_t_Int_t_Int_t)
24,0.9999999999999987,1.0,1.1102230246251565e-13,void_Domain::SetupBoundaryConditions(Int_t)
25,0.9999999999999987,1.0,1.1102230246251565e-13,void_Domain::SetupCommBuffers(Int_t)
26,0.9999999999999987,1.0,1.1102230246251565e-13,void_Domain::SetupElementConnectivities(Int_t)
27,0.9999999999999987,1.0,1.1102230246251565e-13,void_Domain::SetupSymmetryPlanes(Int_t)
28,0.9999999999999987,1.0,1.1102230246251565e-13,void_Domain::~Domain()
29,0.9999999999999987,1.0,1.1102230246251565e-13,void_InitMeshDecomp(Int_t_Int_t_Int_t
30,0.9999999999999987,1.0,1.1102230246251565e-13,void_ParseCommandLineOptions(int_char_**_Int_t_cmdLineOpts
31,0.03220117999999999,0.03220118000000001,8.134579861569993e-05,void_VerifyAndWriteFinalOutput(Real_t_Domain


_tmp_DF.mean(numeric_only=True)

予測値    6.392e+04
実測値    6.392e+04
MAPE       2.638
dtype: float64


_list_series_whole: list[pd.Series] = []
for elem_process in test_lulesh_processes:
    for elem_iteration in test_lulesh_iterations:
        for elem_size in test_lulesh_sizes:
            _list_series: list[pd.Series] = []
            for functionName in functionNames_lulesh:
                target_env_oneCase = [
                    (dict_symbols_lulesh["size"], elem_size),
                    (dict_symbols_lulesh["iteration"], elem_iteration),
                    (dict_symbols_lulesh["process"], elem_process),
                ]

                # コール回数
                _predicted_call: float = (
                    dict_models_lulesh_call[functionName]
                    .subs(target_env_oneCase)
                    .evalf()
                )

                target_lulesh_DF: pd.DataFrame = testDF_lulesh[
                    (testDF_lulesh["Name"] == functionName)
                    & (testDF_lulesh["process"] == elem_process)
                    & (testDF_lulesh["iteration"] == elem_iteration)
                    & (testDF_lulesh["size"] == elem_size)
                ]
                _real_call: float = target_lulesh_DF.reset_index().loc[0]["#Call"]

                _tmp_series: pd.Series = pd.Series(
                    {
                        "Name": functionName,
                        "process": elem_process,
                        "iteration": elem_iteration,
                        "size": elem_size,
                        "_predicted_call": _predicted_call,
                        "_real_call": _real_call,
                    }
                )

                _list_series.append(_tmp_series)
            _tmp_DF: pd.DataFrame = pd.DataFrame(data=_list_series)
            _tmp_DF = _tmp_DF.astype(
                {
                    "process": float,
                    "iteration": float,
                    "size": float,
                    "_predicted_call": float,
                    "_real_call": float,
                }
            )
            _tmp_DF = add_relativeErrorRateCol(
                inputDF=_tmp_DF,
                real_colName="_real_call",
                predicted_colName="_predicted_call",
                targetColName="call relative error rate",
            )
            _tmp_DF["weighted MAPE"] = calcWeightedMAPEscore(
                inputDF=_tmp_DF,
                inputColumnDict={
                    "funcName": "Name",
                    "call": "_real_call",
                    "MAPE": "call relative error rate",
                },
            )
            _list_series_whole.append(_tmp_DF.mean(numeric_only=True))

resultDF_call_prediction_accuracy: pd.DataFrame = pd.DataFrame(data=_list_series_whole)

resultDF_call_prediction_accuracy = resultDF_call_prediction_accuracy.drop(
    columns=["_predicted_call", "_real_call"]
)


print(resultDF_call_prediction_accuracy.to_csv())

,process,iteration,size,call relative error rate,weighted MAPE
0,512.0,256.0,64.0,2.3670808340001734,0.010288561799213546
1,512.0,256.0,96.0,2.3670825851693684,0.0032555301172473203
2,512.0,256.0,128.0,2.367080875286372,0.0014744577828284854
3,512.0,512.0,64.0,2.3669089763021747,0.01019978358968124
4,512.0,512.0,96.0,2.366908992936377,0.0031057606684838095
5,512.0,512.0,128.0,2.3669089969854924,0.0013781898150524797
6,512.0,1024.0,64.0,2.366822902982782,0.01015526465762515
7,512.0,1024.0,96.0,2.366828941775678,0.0032511744607927875
8,512.0,1024.0,128.0,2.3668287284279663,0.0015159890300888623
9,729.0,256.0,64.0,2.5693203049921034,0.011474204365806144
10,729.0,256.0,96.0,2.5693220561612984,0.0036069798468940178
11,729.0,256.0,128.0,2.5693203462783014,0.0016227408673255138
12,729.0,512.0,64.0,2.5691342642882082,0.01138630989535183
13,729.0,512.0,96.0,2.5691342809224107,0.003457472456381772
14,729.0,512.0,128.0,2.569134284971526,0.0015265834769116217
15,729.0,1024.0,64.0,2.5690484715445887,0.011342278758229879
16,729.0,1024.0,96.0,2.5690545103374847,0.00360303089134757
17,729.0,1024.0,128.0,2.569054296989773,0.001664443724473562
18,1000.0,256.0,64.0,2.7254109459355034,0.0124385584670438
19,1000.0,256.0,96.0,2.7254126971046984,0.0038928392832321967
20,1000.0,256.0,128.0,2.725410987221702,0.0017433506488959707
21,1000.0,512.0,64.0,2.7252219764076546,0.0123514286190105
22,1000.0,512.0,96.0,2.7252219930418575,0.0037435586304984067
23,1000.0,512.0,128.0,2.725221997090973,0.0016472889322820992
24,1000.0,1024.0,64.0,2.725116758252199,0.012307669331929958
25,1000.0,1024.0,96.0,2.725122797045095,0.0038891976908803252
26,1000.0,1024.0,128.0,2.7251225836973836,0.0017851832016452954


%reset -f

jupyter_pwd = %pwd
if jupyter_pwd == "/":
    %cd /workspace

# %pdb on

# ipynb形式のライブラリのインポート
%run ./lib/lib.ipynb

# ipynb形式のライブラリノートを.py形式に変更したものをインポート
import lib
import lib.lab_lib
from lib.lab_lib import *

# 生データの入ったCSVファイルの保持されたディレクトリ名を格納している変数
csvDirPath = "./csv_files/"

# NPBのベンチマーク名のリスト
benchmarkNames = ["cg", "ep", "ft", "is", "lu", "mg"]

# NPBのプロセス数
npb_process :list[int] = [2, 4, 8, 16, 32, 64, 128, 256]
train_npb_process :list[int] = npb_process[:-1]
test_npb_process :list[int] = npb_process[-1:]
# NPBのCGの初期変数
cg_na: list[int] = [14000, 30000, 75000, 100000, 1500000]
cg_nonzer: list[int] = [11, 12, 13, 14, 15, 18, 21]
cg_niter: list[int] = [15, 30, 75, 90, 100]
cg_shift: list[int] = [20, 40, 60, 80, 110, 200]

train_cg_na: list[int] = cg_na[:-1]
train_cg_nonzer: list[int] = cg_nonzer[:-1]
train_cg_niter: list[int] = cg_niter[:-1]
train_cg_shift: list[int] = cg_shift[:-1]

test_cg_na: list[int] = cg_na[-1:]
test_cg_nonzer: list[int] = cg_nonzer[-1:]
test_cg_niter: list[int] = cg_niter[-1:]
test_cg_shift: list[int] = cg_shift[-1:]
# NPBのMGの初期変数
mg_size :list[int] = [32, 64, 128, 256, 512]
mg_nit: list[int] = [4, 10, 20, 35, 50]

train_mg_process :list[int] = [4, 8, 16, 32, 64]
train_mg_size :list[int] = [4, 8, 16, 32, 64]
train_mg_nit :list[int] = [5, 10, 15, 20, 25]

test_mg_process :list[int] = [128, 256, 512]
test_mg_size :list[int] = [128, 256, 512]
test_mg_nit :list[int] = [30, 40, 50]

# NPBのEPの初期変数
train_ep_process :list[int] = [4, 8, 16, 32, 64]
train_ep_size :list[str] = ["S", "W", "A", "B", "C"]
test_ep_process :list[int] = [128, 256, 512]
test_ep_size :list[str] = ["D", "E", "F"]

# NPBのFTの初期変数
train_ft_process :list[int] = [2, 4, 8, 16, 32, 64]

# NPBのISの初期変数
train_is_process :list[int] = [2, 4, 8, 16, 32, 64]

# LULESH ベンチマークプログラムのプロセス数・問題サイズ・イテレーション数
lulesh_processes: list[int] = [8, 27, 64, 125, 216, 343, 512]
lulesh_iterations: list[int] = [8, 16, 32, 64, 128, 256]
lulesh_sizes: list[int] = [16, 24, 32, 48, 64, 128]

train_lulesh_processes: list[int] = [8, 27, 64, 125, 216, 343]
train_lulesh_iterations: list[int] = [8, 16, 32, 64, 128]
train_lulesh_sizes: list[int] = [16, 24, 32, 48]

test_lulesh_processes: list[int] = [512, 729, 1000]
test_lulesh_iterations: list[int] = [256, 512, 1024]
test_lulesh_sizes: list[int] = [64, 96, 128]

# Extra-Pのオプション
modelerNames: list[str] = [
    # "refining", 
    "multi-parameter",
    "default", 
    # "basic --options poly_exponents=-1,0,1,2,3 log_exponents=0,1 force_combination_exponents=1 allow_negative_exponents=1"
    ]

modelerOption: str = """ --options \#spm=Basic \#spo=poly_exponents=-1,0,1,2,3,log_exponents=0,1,force_combination_exponents=1,allow_negative_exponents=True"""

list_modelName: list[str] = [
    "modelIp",
    "modelLog",
    "modelLinAndIp",
    "modelLinAndLog",
    "modelIpAndLin",
    "modelIpAndLog",
    "modelLogAndLin",
    "modelLogAndIp",
    "modelProcessDividedByProblemSize",
    "modelProblemSizeDividedByProcess",
    "modelInfiniteProductOfProblemSizeMultipliedByProcesses",
    "modelInfiniteProductOfProblemSizeDividedByProcesses",
    "modelLinearSumOf2elementCombination",
    "modelLinearSumOfElementCombinations",
    "modelLinearSumOf2elementCombinationWithSquared",
    "modelLinearSumOf2elementCombinationWithCubed",
    "modelSquareRootOfProcess",
    "modelSquareRootTimesOtherElems",
    "modelObeyOneParameter",
    "modelLin"
    # "modelBasicTree",
]
list_csvDir = [
    "./csv_files/lulesh_1st/",
    "./csv_files/lulesh_2nd/",
    "./csv_files/lulesh_3rd/",
]

list_csvDir_mg = [
    "./csv_files/mg_1st/",
    "./csv_files/mg_2nd/",
    "./csv_files/mg_3rd/",
]

DEBUG:__main__:hello


_tmp_DF: pd.DataFrame = return_rawDF_mg(
    list_process=[test_mg_process[0]],
    list_problem_size=[test_mg_size[0]],
    list_nit=[test_mg_nit[0]],
    csvDir=list_csvDir_mg[0],
)


trainDF_mg: pd.DataFrame = ret_averaged_rawDF_mg(
    list_process=train_mg_process,
    list_nit=train_mg_nit,
    list_size=train_mg_size,
    list_csvDir=list_csvDir_mg,
    resVar="Exclusive",
)
testDF_mg: pd.DataFrame = ret_averaged_rawDF_mg(
    list_process=test_mg_process,
    list_nit=test_mg_nit,
    list_size=test_mg_size,
    list_csvDir=list_csvDir_mg,
    resVar="Exclusive",
)

trainDF_mg = returnConvertedTargetPprofTimeDF(
    inputDF=trainDF_mg, resVars=["Exclusive", "Inclusive"]
)
testDF_mg = returnConvertedTargetPprofTimeDF(
    inputDF=testDF_mg, resVars=["Exclusive", "Inclusive"]
)


trainDF_mg: pd.DataFrame = return_rawDF_mg(
    list_process=train_npb_process,
    list_problem_size=train_mg_size,
    list_nit=train_mg_nit,
    csvDir=csvDirPath,
)

testDF_mg: pd.DataFrame = return_rawDF_mg(
    list_process=test_npb_process,
    list_nit=test_mg_nit,
    list_problem_size=test_mg_size,
    csvDir=csvDirPath,
)

trainDF_mg = returnConvertedTargetPprofTimeDF(
    inputDF=trainDF_mg, resVars=["Exclusive", "Inclusive"]
)
testDF_mg = returnConvertedTargetPprofTimeDF(
    inputDF=testDF_mg, resVars=["Exclusive", "Inclusive"]
)

trainDF_mg = addPerCallCol(
    inputDF=trainDF_mg, targetColNames=["Exclusive", "Inclusive"], CallColName="#Call"
)
testDF_mg = addPerCallCol(
    inputDF=testDF_mg, targetColNames=["Exclusive", "Inclusive"], CallColName="#Call"
)

/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit5_process2.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit10_process2.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit15_process2.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit20_process2.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit25_process2.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit5_process2.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit10_process2.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit15_process2.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit20_process2.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit25_process2.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit5_process2.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit10_process2.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit15_process2.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit20_process2.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit25_process2.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size32_nit5_process2.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size32_nit15_process2.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size32_nit25_process2.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size64_nit5_process2.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size64_nit15_process2.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size64_nit25_process2.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit5_process4.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit10_process4.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit15_process4.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit20_process4.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit25_process4.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit5_process4.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit10_process4.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit15_process4.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit20_process4.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit25_process4.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit5_process4.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit10_process4.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit15_process4.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit20_process4.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit25_process4.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size32_nit5_process4.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size32_nit15_process4.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size32_nit25_process4.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size64_nit5_process4.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size64_nit15_process4.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size64_nit25_process4.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit5_process8.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit10_process8.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit15_process8.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit20_process8.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit25_process8.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit5_process8.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit10_process8.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit15_process8.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit20_process8.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit25_process8.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit5_process8.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit10_process8.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit15_process8.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit20_process8.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit25_process8.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size32_nit5_process8.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size32_nit15_process8.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size32_nit25_process8.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size64_nit5_process8.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size64_nit15_process8.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size64_nit25_process8.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit5_process16.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit10_process16.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit15_process16.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit20_process16.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit25_process16.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit5_process16.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit10_process16.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit15_process16.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit20_process16.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit25_process16.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit5_process16.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit10_process16.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit15_process16.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit20_process16.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit25_process16.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size32_nit5_process16.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size32_nit15_process16.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size32_nit25_process16.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size64_nit5_process16.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size64_nit15_process16.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size64_nit25_process16.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit5_process32.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit10_process32.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit15_process32.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit20_process32.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit25_process32.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit5_process32.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit10_process32.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit15_process32.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit20_process32.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit25_process32.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit5_process32.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit10_process32.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit15_process32.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit20_process32.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit25_process32.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size32_nit5_process32.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size32_nit15_process32.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size32_nit25_process32.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size64_nit5_process32.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size64_nit15_process32.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size64_nit25_process32.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit5_process64.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit10_process64.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit15_process64.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit20_process64.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit25_process64.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit5_process64.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit10_process64.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit15_process64.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit20_process64.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit25_process64.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit5_process64.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit10_process64.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit15_process64.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit20_process64.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit25_process64.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size32_nit5_process64.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size32_nit15_process64.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size32_nit25_process64.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size64_nit5_process64.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size64_nit15_process64.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size64_nit25_process64.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit5_process128.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit10_process128.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit15_process128.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit20_process128.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size4_nit25_process128.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit5_process128.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit10_process128.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit15_process128.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit20_process128.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size8_nit25_process128.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit5_process128.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit10_process128.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit15_process128.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit20_process128.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size16_nit25_process128.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size32_nit5_process128.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size32_nit15_process128.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size32_nit25_process128.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size64_nit5_process128.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size64_nit15_process128.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size64_nit25_process128.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size128_nit30_process256.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size128_nit40_process256.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size256_nit30_process256.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size256_nit40_process256.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size512_nit30_process256.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")
/workspace/lib/lab_lib.py:8830: UserWarning: ./csv_files/mg_problem_size512_nit40_process256.csv doesn't exist
  warnings.warn(f"{filePath} doesn't exist")


# 学習用のデータ
trainDF_cg: pd.DataFrame = return_rawDF_cg(
    list_process=train_npb_process,
    list_na=train_cg_na,
    list_nonzer=train_cg_nonzer,
    list_niter=train_cg_niter,
    list_shift=train_cg_shift,
    csvDir=csvDirPath,
)

# 予測対象用のデータ
testDF_cg: pd.DataFrame = return_rawDF_cg(
    list_process=test_npb_process,
    list_na=test_cg_na,
    list_nonzer=test_cg_nonzer,
    list_niter=test_cg_niter,
    list_shift=test_cg_shift,
    csvDir=csvDirPath,
)

trainDF_cg = returnConvertedTargetPprofTimeDF(
    inputDF=trainDF_cg, resVars=["Exclusive", "Inclusive"]
)
testDF_cg = returnConvertedTargetPprofTimeDF(
    inputDF=testDF_cg, resVars=["Exclusive", "Inclusive"]
)

trainDF_cg = addPerCallCol(
    inputDF=trainDF_cg, targetColNames=["Exclusive", "Inclusive"], CallColName="#Call"
)
testDF_cg = addPerCallCol(
    inputDF=testDF_cg,
    targetColNames=["Exclusive", "Inclusive"],
    CallColName="#Call",
)

/workspace/lib/lab_lib.py:8785: UserWarning: ./csv_files/cg_na75000_nonzer18_niter30_shift60_process4.csv is empty.
  warnings.warn(f"{filePath} is empty.")
/workspace/lib/lab_lib.py:8785: UserWarning: ./csv_files/cg_na75000_nonzer15_niter90_shift80_process8.csv is empty.
  warnings.warn(f"{filePath} is empty.")


"""
functionNames_cg list[str]:
    ベンチマークプログラムCGで実行された関数の関数名を保持した文字列のリスト

functionNames_mg list[str]:
    ベンチマークプログラムMGで実行された関数の関数名を保持した文字列のリスト

expVar_cg list[str]:
    ベンチマークプログラムCGの説明変数

expVar_mg list[str]:
    ベンチマークプログラムMGの説明変数

resVar_call str:
    関数コール回数の目的変数

resVar_excl str:
    Exclusiveの目的変数(all)

resVar_excl_perCall str:
    Exclusiveの目的変数(perCall)

dict_symbols_cg dict[str, any]:
    ベンチマークプログラムCGのシンボルを保持した辞書

dict_symbols_mg dict[str, any]:
    ベンチマークプログラムMGのシンボルを保持した辞書

benchmarkName_cg str:
    ベンチマーク名CG

benchmarkName_mg str:
    ベンチマーク名MG

modelerName str:
    Extra-Pのモデル名を保持した文字列

modelerOption str:
    Extra-Pのモデルオプションを保持した文字列

"""
functionNames_cg: list[str] = sorted(list(set(trainDF_cg["Name"].to_list())))
functionNames_mg: list[str] = sorted(list(set(trainDF_mg["Name"].to_list())))
expVar_cg: list[str] = ["process", "na", "nonzer", "niter", "shift"]
expVar_mg: list[str] = ["process", "problem_size", "nit"]
resVar_call: str = "#Call"
resVar_excl: str = "Exclusive"
resVar_excl_perCall: str = f"{resVar_excl}PerCall"
dict_symbols_cg: dict[str, any] = {}
dict_symbols_mg: dict[str, any] = {}
benchmarkName_cg: str = "cg"
benchmarkName_mg: str = "mg"
modelerName: str = "multi-parameter"
modelerOption: str = """ --options \#spm=Basic \#spo=poly_exponents=-1,0,1,2,3,log_exponents=0,1,force_combination_exponents=1,allow_negative_exponents=True"""

for elem in expVar_cg:
    dict_symbols_cg[elem] = symbols(elem, real=True)
for elem in expVar_mg:
    dict_symbols_mg[elem] = symbols(elem, real=True)


# MGにおけるモデル構築

"""
dict_mg_model_exclusive_all dict[str, any]:
    Exclusive予測用のモデル@all to all
dict_mg_model_exclusive*_perCall dict[str, any]:
    Exclusive予測用のモデル@perCall
dict_cg_model_call dict[str, any]:
    関数コール回数用のモデル
"""
dict_mg_model_exclusive_all: dict[str, any] = {}
dict_mg_model_exclusive_perCall: dict[str, any] = {}
dict_mg_model_call: dict[str, any] = {}

for functionName in functionNames_mg:

    trainDF_mg_perFunc: pd.DataFrame = trainDF_mg[trainDF_mg["Name"] == functionName]

    mg_model_exclusive_all: str = get_ExtraP_model(
        inputDF_perFunc=trainDF_mg_perFunc,
        expVar=expVar_mg,
        resVar=resVar_excl,
        functionName=functionName,
        dict_symbols=dict_symbols_mg,
        benchmarkName=benchmarkName_mg,
        modelerName=modelerName,
        modelerOption=modelerOption,
    )

    mg_model_exclusive_perCall: str = get_ExtraP_model(
        inputDF_perFunc=trainDF_mg_perFunc,
        expVar=expVar_mg,
        resVar=resVar_excl_perCall,
        functionName=functionName,
        dict_symbols=dict_symbols_mg,
        benchmarkName=benchmarkName_mg,
        modelerName=modelerName,
        modelerOption=modelerOption,
    )

    mg_model_call: str = get_ExtraP_model(
        inputDF_perFunc=trainDF_mg_perFunc,
        expVar=expVar_mg,
        resVar=resVar_call,
        functionName=functionName,
        dict_symbols=dict_symbols_mg,
        benchmarkName=benchmarkName_mg,
        modelerName=modelerName,
        modelerOption=modelerOption,
    )

    dict_mg_model_exclusive_all[functionName] = mg_model_exclusive_all
    dict_mg_model_exclusive_perCall[functionName] = mg_model_exclusive_perCall
    dict_mg_model_call[functionName] = mg_model_call


# MGにおける予測

_tmp_list: list[pd.Series] = []
for i, sr in testDF_mg.iterrows():
    functionName: str = sr["Name"]
    target_env: list[set[any]] = []
    for expVar in expVar_mg:
        target_env.append((dict_symbols_mg[expVar], sr[expVar]))

    """
    _predicted_call:
        予測された関数コール回数
    _predicted_excl_all:
        all to all で予測されたExclusive
    _predicted_excl_perCall:
        関数コール回数を経由して予測されたExclusive
    """
    _predicted_call: float
    _predicted_excl_all: float
    _predicted_excl_perCall: float

    _predicted_call = dict_mg_model_call[functionName].subs(target_env).evalf()
    _predicted_excl_all = (
        dict_mg_model_exclusive_all[functionName].subs(target_env).evalf()
    )
    _predicted_excl_perCall = (
        dict_mg_model_exclusive_perCall[functionName].subs(target_env).evalf()
        * _predicted_call
    )

    _tmp_series: pd.Series = pd.Series(
        {
            "real_Exclusive": sr["Exclusive"],
            "real_call": sr["#Call"],
            "predicted_Exclusive_all": _predicted_excl_all,
            "predicted_Exclusive_perCall": _predicted_excl_perCall,
            "predicted_call": _predicted_call,
            "process": sr["process"],
            "problem_size": sr["problem_size"],
            "functionName": functionName,
        }
    )
    _tmp_list.append(_tmp_series)

print(sr)
DF_result_mg: pd.DataFrame = pd.DataFrame(data=_tmp_list)
DF_result_mg

%Time                      0.0
Exclusive                  0.0
Inclusive                  0.0
#Call                  1.00391
#Subrs                     0.0
Name                TIMER_READ
process                    256
problem_size               512
nit                         50
ExclusivePerCall           0.0
InclusivePerCall           0.0
Name: 40, dtype: object


DF_result_mg = add_relativeErrorRateCol(
    inputDF=DF_result_mg,
    real_colName="real_Exclusive",
    predicted_colName="predicted_Exclusive_all",
    targetColName="vs all",
)
DF_result_mg = add_relativeErrorRateCol(
    inputDF=DF_result_mg,
    real_colName="real_Exclusive",
    predicted_colName="predicted_Exclusive_perCall",
    targetColName="vs perCall",
)

print(DF_result_mg.mean())

DF_result_mg

real_Exclusive                    0.039021
real_call                      1158.280603
predicted_Exclusive_all           0.032440
predicted_Exclusive_perCall       0.025953
predicted_call                 1355.365947
process                         256.000000
problem_size                    298.666667
vs all                           98.361106
vs perCall                       88.672502
dtype: float64

/tmp/ipykernel_15984/2370717254.py:14: FutureWarning: The default value of numeric_only in DataFrame.mean is deprecated. In a future version, it will default to False. In addition, specifying 'numeric_only=None' is deprecated. Select only valid columns or specify the value of numeric_only to silence this warning.
  print(DF_result_mg.mean())


DF_result_mg = DF_result_mg.astype(
    {
        "predicted_Exclusive_all": "float",
        "predicted_Exclusive_perCall": "float",
        "predicted_call": float,
        "vs all": "float",
        "vs perCall": "float",
    }
)

print(DF_result_mg.dtypes)

print(DF_result_mg.to_csv(float_format="%.3f"))

real_Exclusive                 float64
real_call                      float64
predicted_Exclusive_all        float64
predicted_Exclusive_perCall    float64
predicted_call                 float64
process                          int64
problem_size                     int64
functionName                    object
vs all                         float64
vs perCall                     float64
dtype: object
,real_Exclusive,real_call,predicted_Exclusive_all,predicted_Exclusive_perCall,predicted_call,process,problem_size,functionName,vs all,vs perCall
0,0.000,1.000,0.000,0.000,1.000,256,128,.TAU_application,66.471,66.471
1,0.000,1.000,0.000,0.000,1.000,256,128,MG_MPI,13.893,13.893
2,0.538,1.000,0.366,0.366,1.000,256,128,MPI_Init(),31.984,31.984
3,0.522,1.000,0.396,0.396,1.000,256,128,MPI_Finalize(),24.152,24.152
4,0.001,51.000,0.000,0.000,51.000,256,128,MG3P,24.458,22.166
5,0.002,1024.000,0.002,0.002,794.500,256,128,COMM3,4.168,3.832
6,0.026,5066.620,0.012,0.011,4053.939,256,128,MPI_Wait(),53.101,59.431
7,0.002,4862.620,0.002,0.002,3880.247,256,128,TAKE3,9.696,9.718
8,0.003,359.000,0.012,0.008,282.500,256,128,RESID,284.521,168.946
9,0.001,357.000,0.006,0.001,280.500,256,128,PSINV,530.279,24.667
10,0.001,306.000,0.001,0.002,229.500,256,128,RPRJ3,45.751,200.516
11,0.001,306.000,0.004,0.003,229.500,256,128,INTERP,323.878,223.939
12,0.000,306.000,0.000,0.000,229.500,256,128,COMM3_EX,11.064,74.606
13,0.000,204.000,0.000,0.000,121.125,256,128,TAKE3_EX,40.691,41.406
14,0.005,5066.620,0.003,0.002,4053.939,256,128,READY,35.877,57.713
15,0.002,4862.620,0.002,0.002,3880.247,256,128,GIVE3,9.335,9.974
16,0.005,5066.620,0.022,0.005,4053.939,256,128,MPI_Send(),347.634,5.208
17,0.000,2.000,0.001,0.001,2.000,256,128,ZRAN3,133.068,133.068
18,0.003,6.000,0.001,0.001,6.000,256,128,MPI_Barrier(),66.914,66.914
19,0.002,88.000,0.003,0.003,88.000,256,128,MPI_Allreduce(),71.378,71.378
20,0.001,5066.620,0.001,0.001,4053.939,256,128,MPI_Irecv(),18.996,39.543
21,0.001,7.000,0.001,0.001,7.000,256,128,MPI_Bcast(),7.318,7.318
22,0.000,4.000,0.000,0.000,4.000,256,128,NORM2U3,294.064,294.064
23,0.000,523.562,0.000,0.004,594.594,256,128,ZERO3,60.870,997.809
24,0.000,204.000,0.000,0.000,121.125,256,128,GIVE3_EX,42.185,43.590
25,0.000,512.000,0.001,0.000,1939.381,256,128,VRANLC,342.732,241.254
26,0.000,642.844,0.000,0.000,2164.941,256,128,RANDLC,281.879,209.720
27,0.000,310.266,0.000,0.000,250.859,256,128,BUBBLE,14.544,21.307
28,0.000,6.000,0.000,0.000,6.000,256,128,POWER,13.088,13.088
29,0.000,1.000,0.000,0.000,1.000,256,128,GET_ACTIVE_NPROCS,14.441,14.441
30,0.000,2.000,0.000,0.000,2.000,256,128,TIMER_START,6.925,6.925
31,0.000,1.000,0.000,0.000,1.000,256,128,ALLOC_SPACE,23.511,23.511
32,0.000,2.000,0.000,0.000,2.000,256,128,SETUP,437.813,437.813
33,0.000,1.000,0.000,0.000,1.000,256,128,MPI_Reduce(),2.937,2.937
34,0.000,17.000,0.000,0.000,17.000,256,128,TIMER_CLEAR,3.659,3.659
35,0.000,0.004,0.000,0.000,0.004,256,128,CHECK_TIMER_FLAG,131.259,7.027
36,0.000,2.000,0.000,0.000,2.000,256,128,TIMER_STOP,4.800,4.800
37,0.000,0.004,-0.000,0.000,0.004,256,128,PRINT_RESULTS,304.483,53.199
38,0.000,1.000,0.000,0.000,1.000,256,128,MPI_Comm_size(),4.987,4.987
39,0.000,1.000,0.000,0.000,1.000,256,128,MPI_Comm_rank(),19.798,19.798
40,0.000,1.004,0.000,0.000,1.004,256,128,TIMER_READ,26.190,12.057
41,0.000,1.000,0.000,0.000,1.000,256,256,.TAU_application,38.937,38.937
42,0.000,1.000,0.000,0.000,1.000,256,256,MG_MPI,39.597,39.597
43,0.539,1.000,0.366,0.366,1.000,256,256,MPI_Init(),32.110,32.110
44,0.519,1.000,0.396,0.396,1.000,256,256,MPI_Finalize(),23.714,23.714
45,0.001,51.000,0.000,0.000,51.000,256,256,MG3P,31.183,29.095
46,0.002,1177.000,0.002,0.002,794.500,256,256,COMM3,4.168,3.832
47,0.086,5984.620,0.012,0.011,4053.939,256,256,MPI_Wait(),85.821,87.735
48,0.022,410.000,0.092,0.062,282.500,256,256,RESID,319.015,181.308
49,0.004,5780.620,0.002,0.002,3880.247,256,256,TAKE3,45.152,54.859
50,0.011,408.000,0.050,0.002,280.500,256,256,PSINV,354.942,79.161
51,0.004,357.000,0.004,0.008,229.500,256,256,RPRJ3,2.030,109.767
52,0.004,5780.620,0.002,0.002,3880.247,256,256,GIVE3,45.333,54.987
53,0.024,5984.620,0.047,0.005,4053.939,256,256,MPI_Send(),94.848,78.082
54,0.020,5984.620,0.003,0.002,4053.939,256,256,READY,83.969,88.640
55,0.004,357.000,0.015,0.011,229.500,256,256,INTERP,273.133,175.627
56,0.000,357.000,0.000,0.000,229.500,256,256,COMM3_EX,13.695,69.440
57,0.000,204.000,0.000,0.000,121.125,256,256,TAKE3_EX,37.360,38.115
58,0.001,2.000,0.004,0.004,2.000,256,256,ZRAN3,268.028,268.028
59,0.002,5984.620,0.001,0.001,4053.939,256,256,MPI_Irecv(),59.498,30.228
60,0.002,88.000,0.003,0.003,88.000,256,256,MPI_Allreduce(),71.378,71.378
61,0.002,6.000,0.001,0.001,6.000,256,256,MPI_Barrier(),50.371,50.371
62,0.002,574.562,0.000,0.016,594.594,256,256,ZERO3,84.390,711.424
63,0.001,7.000,0.001,0.001,7.000,256,256,MPI_Bcast(),7.318,7.318
64,0.000,4.000,0.002,0.002,4.000,256,256,NORM2U3,386.444,386.444
65,0.001,2048.000,0.002,0.002,7757.525,256,256,VRANLC,259.799,169.691
66,0.000,2.000,0.000,0.000,2.000,256,256,TIMER_START,92.084,92.084
67,0.000,2222.550,0.002,0.001,8426.156,256,256,RANDLC,321.849,241.706
68,0.000,204.000,0.000,0.000,121.125,256,256,GIVE3_EX,39.494,40.965
69,0.000,390.875,0.000,0.000,250.859,256,256,BUBBLE,31.956,37.341
70,0.000,6.000,0.000,0.000,6.000,256,256,POWER,21.872,21.872
71,0.000,1.000,0.000,0.000,1.000,256,256,GET_ACTIVE_NPROCS,15.228,15.228
72,0.000,1.000,0.000,0.000,1.000,256,256,ALLOC_SPACE,28.536,28.536
73,0.000,2.000,0.000,0.000,2.000,256,256,SETUP,342.504,342.504
74,0.000,1.000,0.000,0.000,1.000,256,256,MPI_Reduce(),16.325,16.325
75,0.000,17.000,0.000,0.000,17.000,256,256,TIMER_CLEAR,7.252,7.252
76,0.000,0.004,0.000,0.000,0.004,256,256,CHECK_TIMER_FLAG,142.389,2.553
77,0.000,2.000,0.000,0.000,2.000,256,256,TIMER_STOP,24.183,24.183
78,0.000,0.004,-0.000,0.000,0.004,256,256,PRINT_RESULTS,303.058,53.525
79,0.000,1.000,0.000,0.000,1.000,256,256,MPI_Comm_size(),15.183,15.183
80,0.000,1.000,0.000,0.000,1.000,256,256,MPI_Comm_rank(),10.395,10.395
81,0.000,1.004,0.000,0.000,1.004,256,256,TIMER_READ,29.058,14.604
82,0.000,1.000,0.000,0.000,1.000,256,512,.TAU_application,42.851,42.851
83,0.000,1.000,0.000,0.000,1.000,256,512,MG_MPI,49.934,49.934
84,0.001,51.000,0.000,0.000,51.000,256,512,MG3P,56.714,55.401
85,0.408,461.000,0.737,0.492,282.500,256,512,RESID,80.727,20.697
86,0.003,1330.000,0.002,0.002,794.500,256,512,COMM3,30.555,30.779
87,0.543,1.000,0.366,0.366,1.000,256,512,MPI_Init(),32.610,32.610
88,0.518,1.000,0.396,0.396,1.000,256,512,MPI_Finalize(),23.566,23.566
89,0.144,459.000,0.400,0.008,280.500,256,512,PSINV,177.757,94.133
90,0.031,6698.620,0.002,0.002,3880.247,256,512,GIVE3,92.946,94.192
91,0.222,6902.620,0.096,0.005,4053.939,256,512,MPI_Send(),56.970,97.630
92,0.215,6902.620,0.012,0.011,4053.939,256,512,MPI_Wait(),94.329,95.094
93,0.024,6698.620,0.002,0.002,3880.247,256,512,TAKE3,90.859,92.476
94,0.060,408.000,0.015,0.033,229.500,256,512,RPRJ3,74.252,44.322
95,0.113,408.000,0.060,0.044,229.500,256,512,INTERP,47.148,61.296
96,0.103,6902.620,0.003,0.003,4053.939,256,512,READY,96.887,97.488
97,0.000,408.000,0.000,0.000,229.500,256,512,COMM3_EX,39.479,18.819
98,0.007,2.000,0.015,0.015,2.000,256,512,ZRAN3,108.406,108.406
99,0.000,204.000,0.000,0.000,121.125,256,512,TAKE3_EX,32.585,33.398
100,0.014,625.562,0.001,0.065,594.594,256,512,ZERO3,92.964,364.161
101,0.004,4.000,0.014,0.014,4.000,256,512,NORM2U3,254.718,254.718
102,0.006,6.000,0.001,0.001,6.000,256,512,MPI_Barrier(),83.457,83.457
103,0.004,6902.620,0.001,0.001,4053.939,256,512,MPI_Irecv(),79.749,65.114
104,0.003,88.000,0.003,0.003,88.000,256,512,MPI_Allreduce(),14.252,14.252
105,0.003,8192.000,0.008,0.006,31030.101,256,512,VRANLC,178.711,107.482
106,0.001,7.000,0.001,0.001,7.000,256,512,MPI_Bcast(),7.318,7.318
107,0.001,8442.270,0.006,0.005,33471.013,256,512,RANDLC,538.650,417.150
108,0.001,1.000,0.000,0.000,1.000,256,512,ALLOC_SPACE,98.042,98.042
109,0.000,204.000,0.000,0.000,121.125,256,512,GIVE3_EX,35.491,37.059
110,0.000,2.000,0.000,0.000,2.000,256,512,TIMER_START,78.763,78.763
111,0.000,471.828,0.000,0.000,250.859,256,512,BUBBLE,40.422,45.136
112,0.000,6.000,0.000,0.000,6.000,256,512,POWER,36.139,36.139
113,0.000,1.000,0.000,0.000,1.000,256,512,GET_ACTIVE_NPROCS,4.375,4.375
114,0.000,2.000,0.000,0.000,2.000,256,512,SETUP,388.920,388.920
115,0.000,1.000,0.000,0.000,1.000,256,512,MPI_Reduce(),15.155,15.155
116,0.000,17.000,0.000,0.000,17.000,256,512,TIMER_CLEAR,15.553,15.553
117,0.000,0.004,0.000,0.000,0.004,256,512,CHECK_TIMER_FLAG,131.259,7.027
118,0.000,2.000,0.000,0.000,2.000,256,512,TIMER_STOP,39.347,39.347
119,0.000,0.004,-0.000,0.000,0.004,256,512,PRINT_RESULTS,311.534,51.585
120,0.000,1.000,0.000,0.000,1.000,256,512,MPI_Comm_size(),23.983,23.983
121,0.000,1.000,0.000,0.000,1.000,256,512,MPI_Comm_rank(),13.217,13.217
122,0.000,1.004,0.000,0.000,1.004,256,512,TIMER_READ,49.436,32.699

	real_Exclusive	functionName	predicted_Exclusive	predicted_Exclusive_from_call	predicted_call	real_call	real vs predicted all @Exclusive	real vs predicted using call @Exclusive
0	3.366667e-06	.TAU_application	3.58708333333333e-6	2.10570898355123e-6	0.999999999999999	3.366667e-06	6.54702970297031	37.4541886073892
1	5.133010e+02	MPI_Allreduce()	437.915990349888	12124.7559208733	69114.4138888889	5.133010e+02	14.6863165374920	2262.11422165031
2	3.833333e-02	MPI_Barrier()	0.000816330555555555	18.5004563418933	662.485549774719	3.833333e-02	97.8704420289855	48162.0600223303
3	9.000000e-03	MPI_Comm_rank()	0.0165749444775296	0.00575717034518276	7682.31799535170	9.000000e-03	84.1660497503293	36.0314406090804
4	4.553333e-07	MPI_Comm_size()	4.85952941689757e-7	4.90953764193764e-7	0.988953189305555	4.553333e-07	6.72465776495400	7.82293503523359
5	7.433333e-01	MPI_Finalize()	0.488460112539233	130774.528268482	32774.5344861112	7.433333e-01	34.2878772368745	17592885.8657150
6	6.153333e-01	MPI_Init()	0.493019444444445	15072.7188036146	30426.8863055556	6.153333e-01	19.8776634163958	2449420.93233174
7	5.633333e-02	MPI_Irecv()	0.0607189920652085	-0.0896068042327797	-37413.2240107461	5.633333e-02	7.78519301516302	259.065332957597
8	8.950000e-01	MPI_Isend()	3.17498669266321	1.11110221315045	26027.4814559232	8.950000e-01	254.747116498683	24.1454986760275
9	7.166667e-02	MPI_Reduce()	0.0581820601722947	786.657959823452	16456.6172194788	7.166667e-02	18.8157299921470	1097562.26952110
10	4.289367e+01	MPI_Wait()	71.5007503928785	11.6430685704400	5079.97374444445	4.289367e+01	66.6930247500683	72.8559727455335
11	1.180570e+02	MPI_Waitall()	1919.56884083925	159223.010844645	1975772.85318665	1.180570e+02	1525.96782981038	134769.606075578
12	3.164563e+02	Real_t_CalcElemVolume(const	335.378503734694	293.728744937127	2030468957.63812	3.164563e+02	5.97939380831707	7.18190347363545
13	2.020000e-06	StrToInt	1.91816666666667e-6	2.17858055555555e-6	2.29777777777778	2.020000e-06	5.04125412541255	7.85052255225513
14	3.181085e+03	int_main(int_char_**)	2135.89439042910	2103.48508973343	0.999999999999999	3.181085e+03	32.8564187870145	33.8752315724532
15	4.528560e+02	void_CalcKinematicsForElems(Domain	445.419547861958	668.181607918078	1526.19624444444	4.528560e+02	1.64212291281154	47.5483614919706
16	1.440000e+00	void_CommMonoQ(Domain	5.25854310710730	25.1601639450549	8809.01525702766	1.440000e+00	265.176604660229	1647.23360729548
17	4.700000e-02	void_CommRecv(Domain	0.00195521666666667	-0.367699587925509	3509.64698308692	4.700000e-02	95.8399645390071	882.339548777678
18	2.805667e+00	void_CommSBN(Domain	1.57211617935831	1.70392941987339	366.193383333333	2.805667e+00	43.9663949379240	39.2682872802640
19	8.500000e+00	void_CommSend(Domain	3.13405379446953	432.176530498747	81149.4056222222	8.500000e+00	63.1287788885937	4984.42977057350
20	1.143733e+01	void_CommSyncPosVel(Domain	21.7678534923438	90.8365082238611	6030.22978611111	1.143733e+01	90.3228039083456	694.210552202096
21	5.866667e-02	void_Domain::BuildMesh(Int_t_Int_t_Int_t)	0.0101294785767843	0.385413936939769	43.8605583333333	5.866667e-02	82.7338433350268	556.955574329152
22	2.166667e-02	void_Domain::CreateRegionIndexSets(Int_t_Int_t)	0.00722072107514937	0.0574410726950169	7.60888888888891	2.166667e-02	66.6735950377721	165.112643207770
23	5.293333e-01	void_Domain::Domain(Int_t_Index_t_Index_t_Inde...	0.731860884361024	283.485769620082	293.553219444444	5.293333e-01	38.2608723603950	53455.2461498896
24	7.000000e-03	void_Domain::SetupBoundaryConditions(Int_t)	0.00485935192872183	0.0233812701063779	5.05637777777779	7.000000e-03	30.5806867325452	234.018144376826
25	4.000000e-03	void_Domain::SetupCommBuffers(Int_t)	0.00486234543933642	0.0972049848525025	19.5226333333333	4.000000e-03	21.5586359834105	2330.12462131256
26	1.866667e-02	void_Domain::SetupElementConnectivities(Int_t)	0.0171993242439927	2.23505193201573	130.968395555556	1.866667e-02	7.86076297861032	11873.4924929414
27	2.836667e-05	void_Domain::SetupSymmetryPlanes(Int_t)	3.86090468625521e-5	3.88943505553212e-5	1.01111111111111	2.836667e-05	36.1070982228629	37.1128691726950
28	1.000000e-03	void_Domain::~Domain()	0.000432445444444444	0.000790691726770833	1.85837673611111	1.000000e-03	56.7554555555556	20.9308273229167
29	1.226667e-05	void_InitMeshDecomp(Int_t_Int_t_Int_t	1.12770555555556e-5	1.08301957733312e-5	0.960776748972221	1.226667e-05	8.06748188405792	11.7103605434956
30	1.573333e-06	void_ParseCommandLineOptions(int_char_**_Int_t...	2.08527777777778e-6	2.01648854425741e-6	0.978600822222221	1.573333e-06	32.5388418079097	28.1666447621233
31	1.296667e-07	void_VerifyAndWriteFinalOutput(Real_t_Domain	9.57197219605595e-8	9.60599283426180e-6	0.109382572277778	1.296667e-07	26.1801630124220	7308.22069480344

	real_Inclusive	functionName	predicted_Inclusive	predicted_Inclusive_from_call	predicted_call	real_call	real vs predicted all @Inclusive	real vs predicted using call @Inclusive
0	4.651960e+03	.TAU_application	4112.13002651492	3755.07819771439	0.999999999999999	4.651960e+03	11.6043617773413	19.2796600003743
1	5.133010e+02	MPI_Allreduce()	439.345968776950	12124.7559208733	69114.4138888889	5.133010e+02	14.4077317642182	2262.11422165031
2	3.833333e-02	MPI_Barrier()	0.000869176944444444	18.5004563418933	662.485549774719	3.833333e-02	97.7325818840580	48162.0600223303
3	9.000000e-03	MPI_Comm_rank()	0.0165754414998621	0.00575717034518276	7682.31799535170	9.000000e-03	84.1715722206905	36.0314406090804
4	4.553333e-07	MPI_Comm_size()	4.85952941689757e-7	4.90953764193764e-7	0.988953189305555	4.553333e-07	6.72465776495400	7.82293503523359
5	7.433333e-01	MPI_Finalize()	0.501003301920031	130774.528268482	32774.5344861112	7.433333e-01	32.6004526564981	17592885.8657150
6	6.153333e-01	MPI_Init()	0.503405555555556	15072.7188036146	30426.8863055556	6.153333e-01	18.1897797038642	2449420.93233174
7	8.066667e-02	MPI_Irecv()	0.0635597911803491	-0.0896068042327797	-37413.2240107461	8.066667e-02	21.2068704375837	211.082815164603
8	8.950000e-01	MPI_Isend()	0.426386241911113	1.11110221315045	26027.4814559232	8.950000e-01	52.3590791160767	24.1454986760275
9	7.166667e-02	MPI_Reduce()	0.0600932981444429	786.657959823452	16456.6172194788	7.166667e-02	16.1488863100797	1097562.26952110
10	4.289367e+01	MPI_Wait()	72.0485507261410	11.6430685704400	5079.97374444445	4.289367e+01	67.9701371441183	72.8559727455335
11	1.180570e+02	MPI_Waitall()	2087.96064418750	159223.010844645	1975772.85318665	1.180570e+02	1668.60384745292	134769.606075578
12	3.164563e+02	Real_t_CalcElemVolume(const	337.081473253762	293.728744937127	2030468957.63812	3.164563e+02	6.51753109289281	7.18190347363545
13	2.020000e-06	StrToInt	1.91816666666667e-6	2.17858055555555e-6	2.29777777777778	2.020000e-06	5.04125412541255	7.85052255225513
14	4.651960e+03	int_main(int_char_**)	4112.13124454251	3755.07819771439	0.999999999999999	4.651960e+03	11.6043355942377	19.2796600003743
15	7.690063e+02	void_CalcKinematicsForElems(Domain	748.661450811180	1138.63521296804	1526.19624444444	7.690063e+02	2.64560662770699	48.0657783444397
16	5.460667e+00	void_CommMonoQ(Domain	58.6305977858969	3.76506625783785	8809.01525702766	5.460667e+00	973.689374665429	31.0511611920794
17	9.966667e-02	void_CommRecv(Domain	0.0979398580451694	0.0957780183755544	3509.64698308692	9.966667e-02	1.73258390116784	3.90165380379157
18	2.321700e+01	void_CommSBN(Domain	0.590025671762993	0.256589508063414	366.193383333333	2.321700e+01	97.4586480950898	98.8948205708601
19	1.274573e+02	void_CommSend(Domain	319.835497015011	7305.78070596617	81149.4056222222	1.274573e+02	150.935343342356	5631.94222325340
20	2.990133e+01	void_CommSyncPosVel(Domain	40.4090521433832	274.587516028816	6030.22978611111	2.990133e+01	35.1413052150957	818.311946051958
21	5.866667e-02	void_Domain::BuildMesh(Int_t_Int_t_Int_t)	0.0102553923824921	0.385413936939769	43.8605583333333	5.866667e-02	82.5192175298429	556.955574329152
22	2.166667e-02	void_Domain::CreateRegionIndexSets(Int_t_Int_t)	0.00763415378183217	0.0574109414347394	7.60888888888891	2.166667e-02	64.7654440838515	164.973575852643
23	9.360000e-01	void_Domain::Domain(Int_t_Index_t_Index_t_Inde...	1.30391133111838	441.271000398884	293.553219444444	9.360000e-01	39.3067661451264	47044.3376494534
24	7.000000e-03	void_Domain::SetupBoundaryConditions(Int_t)	0.00485935192872183	0.0233812701063779	5.05637777777779	7.000000e-03	30.5806867325452	234.018144376826
25	4.000000e-03	void_Domain::SetupCommBuffers(Int_t)	0.00504001017854632	0.0972049848525025	19.5226333333333	4.000000e-03	26.0002544636580	2330.12462131256
26	1.866667e-02	void_Domain::SetupElementConnectivities(Int_t)	0.0171686507686722	2.23505193201573	130.968395555556	1.866667e-02	8.02508516782757	11873.4924929414
27	2.836667e-05	void_Domain::SetupSymmetryPlanes(Int_t)	3.88307804958058e-5	3.88943505553212e-5	1.01111111111111	2.836667e-05	36.8887679053082	37.1128691726950
28	1.000000e-03	void_Domain::~Domain()	0.000432445444444444	0.000790691726770833	1.85837673611111	1.000000e-03	56.7554555555556	20.9308273229167
29	1.226667e-05	void_InitMeshDecomp(Int_t_Int_t_Int_t	1.12770555555556e-5	1.08301957733312e-5	0.960776748972221	1.226667e-05	8.06748188405792	11.7103605434956
30	3.593333e-06	void_ParseCommandLineOptions(int_char_**_Int_t...	3.94702777777778e-6	4.73434483566820e-6	0.978600822222221	3.593333e-06	9.84307359307357	31.7535668553303
31	1.296667e-07	void_VerifyAndWriteFinalOutput(Real_t_Domain	1.36824327345807e-7	9.60599283426180e-6	0.109382572277778	1.296667e-07	5.52004679625255	7308.22069480344

	real_#Subrs	functionName	predicted_#Subrs	predicted_#Subrs_from_call	predicted_call	real_call	real vs predicted all @#Subrs	real vs predicted using call @#Subrs
0	1.000000e+00	.TAU_application	0.999999999999999	0.999999999999998	0.999999999999999	1.000000e+00	1.11022302462516e-13	2.22044604925031e-13
1	0.000000e+00	MPI_Allreduce()	1012881.74466418	0	69114.4138888889	0.000000e+00	zoo	nan
2	0.000000e+00	MPI_Barrier()	-12.4034790238095	0	662.485549774719	0.000000e+00	zoo	nan
3	0.000000e+00	MPI_Comm_rank()	0.00833333333333333	0	7682.31799535170	0.000000e+00	zoo	nan
4	0.000000e+00	MPI_Comm_size()	0	0	0.988953189305555	0.000000e+00	nan	nan
5	0.000000e+00	MPI_Finalize()	10781.0178333333	0	32774.5344861112	0.000000e+00	zoo	nan
6	0.000000e+00	MPI_Init()	10802.6457222222	0	30426.8863055556	0.000000e+00	zoo	nan
7	1.360193e+04	MPI_Irecv()	609.931260790102	0	-37413.2240107461	1.360193e+04	95.5158487705907	100.000000000000
8	0.000000e+00	MPI_Isend()	27108.9717494639	0	26027.4814559232	0.000000e+00	zoo	nan
9	0.000000e+00	MPI_Reduce()	57.6759273283490	0	16456.6172194788	0.000000e+00	zoo	nan
10	0.000000e+00	MPI_Wait()	2561.70743888889	0	5079.97374444445	0.000000e+00	zoo	nan
11	0.000000e+00	MPI_Waitall()	61716.8215638889	0	1975772.85318665	0.000000e+00	zoo	nan
12	0.000000e+00	Real_t_CalcElemVolume(const	2133.78080555556	0	2030468957.63812	0.000000e+00	zoo	nan
13	0.000000e+00	StrToInt	0	0	2.29777777777778	0.000000e+00	nan	nan
14	1.127600e+04	int_main(int_char_**)	11276.0268521505	11276.0268521505	0.999999999999999	1.127600e+04	0.000238135425138490	0.000238135425025569
15	2.147480e+09	void_CalcKinematicsForElems(Domain	1926794448.76651	3200665305.73276	1526.19624444444	2.147480e+09	10.2764892447655	49.0428458347813
16	6.553600e+03	void_CommMonoQ(Domain	393545.089707081	47631.6026146341	8809.01525702766	6.553600e+03	5905.02151042299	626.800577005525
17	2.720387e+04	void_CommRecv(Domain	149067.171059722	34082.2257910500	3509.64698308692	2.720387e+04	447.963173346886	25.2844906522479
18	2.250080e+04	void_CommSBN(Domain	94859.2759238944	5667.74006367034	366.193383333333	2.250080e+04	321.581792309137	74.8109397724955
19	4.387880e+04	void_CommSend(Domain	18973116.3199341	869192.097502817	81149.4056222222	4.387880e+04	43139.8249722739	1880.89304516718
20	1.175140e+04	void_CommSyncPosVel(Domain	13265.3614213333	49681.4284501634	6030.22978611111	1.175140e+04	12.8832430291993	322.770295030067
21	0.000000e+00	void_Domain::BuildMesh(Int_t_Int_t_Int_t)	363.131579359959	0	43.8605583333333	0.000000e+00	zoo	nan
22	1.000000e+00	void_Domain::CreateRegionIndexSets(Int_t_Int_t)	67272.7613546246	7.60888888888891	7.60888888888891	1.000000e+00	6727176.13546246	660.888888888891
23	2.097160e+06	void_Domain::Domain(Int_t_Index_t_Index_t_Inde...	1629947.11989493	615627482.583672	293.553219444444	2.097160e+06	22.2783612173164	29255.2939491346
24	0.000000e+00	void_Domain::SetupBoundaryConditions(Int_t)	0	0	5.05637777777779	0.000000e+00	nan	nan
25	0.000000e+00	void_Domain::SetupCommBuffers(Int_t)	6.06878055555556	0	19.5226333333333	0.000000e+00	zoo	nan
26	0.000000e+00	void_Domain::SetupElementConnectivities(Int_t)	4.64711666666667	0	130.968395555556	0.000000e+00	zoo	nan
27	0.000000e+00	void_Domain::SetupSymmetryPlanes(Int_t)	0.0111111111111111	0	1.01111111111111	0.000000e+00	zoo	nan
28	0.000000e+00	void_Domain::~Domain()	0	0	1.85837673611111	0.000000e+00	nan	nan
29	0.000000e+00	void_InitMeshDecomp(Int_t_Int_t_Int_t	0	0	0.960776748972221	0.000000e+00	nan	nan
30	2.000000e+00	void_ParseCommandLineOptions(int_char_**_Int_t...	1.94444444444444	1.95720164444444	0.978600822222221	2.000000e+00	2.77777777777788	2.13991777777799
31	0.000000e+00	void_VerifyAndWriteFinalOutput(Real_t_Domain	0.0444444444444444	0	0.109382572277778	0.000000e+00	zoo	nan

	Name	all to all	perCall to all	real exclusive	real call	vs all to all	vs perCall to all
0	.TAU_application	3.58708333333333e-6	2.10570898355123e-6	3.366667e-06	1.000000e+00	6.54702970297031	37.4541886073892
1	MPI_Allreduce()	437.915990349888	179.465101548773	5.133010e+02	1.023000e+03	14.6863165374920	65.0370637211358
2	MPI_Barrier()	0.000816330555555555	0.0279258262284882	3.833333e-02	1.000000e+00	97.8704420289855	27.1500185343787
3	MPI_Comm_rank()	0.0165749444775296	0.00691026689926807	9.000000e-03	9.221000e+03	84.1660497503293	23.2192566747993
4	MPI_Comm_size()	4.85952941689757e-7	4.96437818799606e-7	4.553333e-07	1.000000e+00	6.72465776495400	9.02733941426189
5	MPI_Finalize()	0.488460112539233	3.99012618543521	7.433333e-01	1.000000e+00	34.2878772368745	436.788276067517
6	MPI_Init()	0.493019444444445	0.495375000000000	6.153333e-01	1.000000e+00	19.8776634163958	19.4948537378115
7	MPI_Irecv()	0.0607189920652085	0.0641030960977833	5.633333e-02	3.773280e+04	7.78519301516302	13.7924782800888
8	MPI_Isend()	3.17498669266321	1.14257579197692	8.950000e-01	3.773280e+04	254.747116498683	27.6620996622256
9	MPI_Reduce()	0.0581820601722947	0.0478019236476088	7.166667e-02	1.000000e+00	18.8157299921470	33.2996414219412
10	MPI_Wait()	71.5007503928785	61.3435829940378	4.289367e+01	3.773280e+04	66.6930247500683	43.0131480032898
11	MPI_Waitall()	1919.56884083925	247.646034581573	1.180570e+02	3.073000e+03	1525.96782981038	109.768192128864
12	Real_t_CalcElemVolume(const	335.378503734694	310.960204792449	3.164563e+02	2.149580e+09	5.97939380831707	1.73677312221611
13	StrToInt	1.91816666666667e-6	1.89625000000000e-6	2.020000e-06	2.000000e+00	5.04125412541255	6.12623762376250
14	int_main(int_char_**)	2135.89439042910	2103.48508973343	3.181085e+03	1.000000e+00	32.8564187870145	33.8752315724532
15	void_CalcKinematicsForElems(Domain	445.419547861958	448.315849943123	4.528560e+02	1.024000e+03	1.64212291281154	1.00255932501224
16	void_CommMonoQ(Domain	5.25854310710730	2.92473189431500	1.440000e+00	1.024000e+03	265.176604660229	103.106381549652
17	void_CommRecv(Domain	0.00195521666666667	-0.321952845725027	4.700000e-02	3.073000e+03	95.8399645390071	785.006054734100
18	void_CommSBN(Domain	1.57211617935831	4.76941347075195	2.805667e+00	1.025000e+03	43.9663949379240	69.9921636242822
19	void_CommSend(Domain	3.13405379446953	16.3658435701341	8.500000e+00	3.073000e+03	63.1287788885937	92.5393361192247
20	void_CommSyncPosVel(Domain	21.7678534923438	15.4250480861394	1.143733e+01	1.024000e+03	90.3228039083456	34.8657736605796
21	void_Domain::BuildMesh(Int_t_Int_t_Int_t)	0.0101294785767843	0.00878725560241810	5.866667e-02	1.000000e+00	82.7338433350268	85.0217234049691
22	void_Domain::CreateRegionIndexSets(Int_t_Int_t)	0.00722072107514937	0.00754920639975394	2.166667e-02	1.000000e+00	66.6735950377721	65.1575089242126
23	void_Domain::Domain(Int_t_Index_t_Index_t_Inde...	0.731860884361024	0.965704856368412	5.293333e-01	1.000000e+00	38.2608723603950	82.4379451577605
24	void_Domain::SetupBoundaryConditions(Int_t)	0.00485935192872183	0.00462411456065167	7.000000e-03	1.000000e+00	30.5806867325452	33.9412205621190
25	void_Domain::SetupCommBuffers(Int_t)	0.00486234543933642	0.00497909186700407	4.000000e-03	1.000000e+00	21.5586359834105	24.4772966751018
26	void_Domain::SetupElementConnectivities(Int_t)	0.0171993242439927	0.0170655822920854	1.866667e-02	1.000000e+00	7.86076297861032	8.57723772097084
27	void_Domain::SetupSymmetryPlanes(Int_t)	3.86090468625521e-5	3.84669401096583e-5	2.836667e-05	1.000000e+00	36.1070982228629	35.6061343466214
28	void_Domain::~Domain()	0.000432445444444444	0.000425474400000000	1.000000e-03	1.000000e+00	56.7554555555556	57.4525600000000
29	void_InitMeshDecomp(Int_t_Int_t_Int_t	1.12770555555556e-5	1.12723333333333e-5	1.226667e-05	1.000000e+00	8.06748188405792	8.10597826086959
30	void_ParseCommandLineOptions(int_char_**_Int_t...	2.08527777777778e-6	2.06058333333333e-6	1.573333e-06	1.000000e+00	32.5388418079097	30.9692796610169
31	void_VerifyAndWriteFinalOutput(Real_t_Domain	9.57197219605595e-8	8.78196087913656e-8	1.296667e-07	1.000000e-03	26.1801630124220	32.2727952765818

	Name	all to all	perCall to all	real inclusive	real call	vs all to all	vs perCall to all
0	.TAU_application	4112.13002651492	3755.07819771439	4.651960e+03	1.000000e+00	11.6043617773413	19.2796600003743
1	MPI_Allreduce()	439.345968776950	179.465101548773	5.133010e+02	1.023000e+03	14.4077317642182	65.0370637211358
2	MPI_Barrier()	0.000869176944444444	0.0279258262284882	3.833333e-02	1.000000e+00	97.7325818840580	27.1500185343787
3	MPI_Comm_rank()	0.0165754414998621	0.00691026689926807	9.000000e-03	9.221000e+03	84.1715722206905	23.2192566747993
4	MPI_Comm_size()	4.85952941689757e-7	4.96437818799606e-7	4.553333e-07	1.000000e+00	6.72465776495400	9.02733941426189
5	MPI_Finalize()	0.501003301920031	3.99012618543521	7.433333e-01	1.000000e+00	32.6004526564981	436.788276067517
6	MPI_Init()	0.503405555555556	0.495375000000000	6.153333e-01	1.000000e+00	18.1897797038642	19.4948537378115
7	MPI_Irecv()	0.0635597911803491	0.0641030960977833	8.066667e-02	3.773280e+04	21.2068704375837	20.5333519448967
8	MPI_Isend()	0.426386241911113	1.14257579197692	8.950000e-01	3.773280e+04	52.3590791160767	27.6620996622256
9	MPI_Reduce()	0.0600932981444429	0.0478019236476088	7.166667e-02	1.000000e+00	16.1488863100797	33.2996414219412
10	MPI_Wait()	72.0485507261410	61.3435829940378	4.289367e+01	3.773280e+04	67.9701371441183	43.0131480032898
11	MPI_Waitall()	2087.96064418750	247.646034581573	1.180570e+02	3.073000e+03	1668.60384745292	109.768192128864
12	Real_t_CalcElemVolume(const	337.081473253762	310.960204792449	3.164563e+02	2.149580e+09	6.51753109289281	1.73677312221611
13	StrToInt	1.91816666666667e-6	1.89625000000000e-6	2.020000e-06	2.000000e+00	5.04125412541255	6.12623762376250
14	int_main(int_char_**)	4112.13124454251	3755.07819771439	4.651960e+03	1.000000e+00	11.6043355942377	19.2796600003743
15	void_CalcKinematicsForElems(Domain	748.661450811180	763.966273880916	7.690063e+02	1.024000e+03	2.64560662770699	0.655398952381933
16	void_CommMonoQ(Domain	58.6305977858969	0.437668426666667	5.460667e+00	1.024000e+03	973.689374665429	91.9850733732145
17	void_CommRecv(Domain	0.0979398580451694	0.0838619530358588	9.966667e-02	3.073000e+03	1.73258390116784	15.8575722048239
18	void_CommSBN(Domain	0.590025671762993	0.718211354260315	2.321700e+01	1.025000e+03	97.4586480950898	96.9065281722000
19	void_CommSend(Domain	319.835497015011	276.658392471159	1.274573e+02	3.073000e+03	150.935343342356	117.059611429047
20	void_CommSyncPosVel(Domain	40.4090521433832	46.6280102726962	2.990133e+01	1.024000e+03	35.1413052150957	55.9395688242316
21	void_Domain::BuildMesh(Int_t_Int_t_Int_t)	0.0102553923824921	0.00878725560241810	5.866667e-02	1.000000e+00	82.5192175298429	85.0217234049691
22	void_Domain::CreateRegionIndexSets(Int_t_Int_t)	0.00763415378183217	0.00754524639183196	2.166667e-02	1.000000e+00	64.7654440838515	65.1757858838525
23	void_Domain::Domain(Int_t_Index_t_Index_t_Inde...	1.30391133111838	1.50320613493525	9.360000e-01	1.000000e+00	39.3067661451264	60.5989460400911
24	void_Domain::SetupBoundaryConditions(Int_t)	0.00485935192872183	0.00462411456065167	7.000000e-03	1.000000e+00	30.5806867325452	33.9412205621190
25	void_Domain::SetupCommBuffers(Int_t)	0.00504001017854632	0.00497909186700407	4.000000e-03	1.000000e+00	26.0002544636580	24.4772966751018
26	void_Domain::SetupElementConnectivities(Int_t)	0.0171686507686722	0.0170655822920854	1.866667e-02	1.000000e+00	8.02508516782757	8.57723772097084
27	void_Domain::SetupSymmetryPlanes(Int_t)	3.88307804958058e-5	3.84669401096583e-5	2.836667e-05	1.000000e+00	36.8887679053082	35.6061343466214
28	void_Domain::~Domain()	0.000432445444444444	0.000425474400000000	1.000000e-03	1.000000e+00	56.7554555555556	57.4525600000000
29	void_InitMeshDecomp(Int_t_Int_t_Int_t	1.12770555555556e-5	1.12723333333333e-5	1.226667e-05	1.000000e+00	8.06748188405792	8.10597826086959
30	void_ParseCommandLineOptions(int_char_**_Int_t...	3.94702777777778e-6	4.83787130376345e-6	3.593333e-06	1.000000e+00	9.84307359307357	34.6346373960143
31	void_VerifyAndWriteFinalOutput(Real_t_Domain	1.36824327345807e-7	8.78196087913656e-8	1.296667e-07	1.000000e-03	5.52004679625255	32.2727952765818

TODO¶

Inclusive¶

排他的な時間予測のモデル適合度¶

包含的な時間予測モデルの適合度¶

関数名を横軸、MAPEを縦軸に取ったグラフ¶

関数コール回数¶

	取得対象の実行規模の数	コスト	凡例
0	1	6.656000e+04	実測
1	2	3.072000e+05	実測
2	3	8.704000e+05	実測
3	4	1.005227e+06	実測
4	5	1.493333e+06	実測
5	6	2.670933e+06	実測
6	7	2.945707e+06	実測
7	8	3.918507e+06	実測
8	9	6.256640e+06	実測
9	10	6.361130e+06	実測
10	11	6.706190e+06	実測
11	12	7.508090e+06	実測
12	13	7.707350e+06	実測
13	14	8.407190e+06	実測
14	15	1.008389e+07	実測
15	16	1.047755e+07	実測
16	17	1.186265e+07	実測
17	18	1.521605e+07	実測
18	19	1.535272e+07	実測
19	20	1.583938e+07	実測
20	21	1.700605e+07	実測
21	22	1.728272e+07	実測
22	23	1.826938e+07	実測
23	24	2.056938e+07	実測
24	25	2.112938e+07	実測
25	26	2.306272e+07	実測
26	27	2.772938e+07	実測
27	0	1.803971e+05	予測
28	1	1.803971e+05	予測
29	2	1.803971e+05	予測
30	3	1.803971e+05	予測
31	4	1.803971e+05	予測
32	5	1.803971e+05	予測
33	6	1.803971e+05	予測
34	7	1.803971e+05	予測
35	8	1.803971e+05	予測
36	9	1.803971e+05	予測
37	10	1.803971e+05	予測
38	11	1.803971e+05	予測
39	12	1.803971e+05	予測
40	13	1.803971e+05	予測
41	14	1.803971e+05	予測
42	15	1.803971e+05	予測
43	16	1.803971e+05	予測
44	17	1.803971e+05	予測
45	18	1.803971e+05	予測
46	19	1.803971e+05	予測
47	20	1.803971e+05	予測
48	21	1.803971e+05	予測
49	22	1.803971e+05	予測
50	23	1.803971e+05	予測
51	24	1.803971e+05	予測
52	25	1.803971e+05	予測
53	27	1.803971e+05	予測

	process	iteration	size	ex_directly	ex_indirectly	ex_real	in_directly	in_indirectly	in_real	vs directly(ex)	vs indirectly(ex)	vs directly(in)	vs indirectly(in)
0	512.0	256.0	64.0	3.919365	2139.335882	0.004130	10.611352	2175.449380	10.992604	93979.761771	3.751343e+08	36.387339	3.916221e+05
1	512.0	256.0	96.0	13.114372	3091.049492	0.014749	35.476425	3210.695513	36.435679	105205.476405	3.755681e+08	36.138631	3.789619e+05
2	512.0	256.0	128.0	31.020158	7377.355767	0.035744	83.877708	7659.679401	86.002127	122737.065763	3.303384e+08	51.632062	3.335468e+05
3	512.0	512.0	64.0	8.713426	3021.576285	0.008229	22.158231	3063.077281	21.919510	98020.774826	7.383209e+08	48.655122	6.237561e+05
4	512.0	512.0	96.0	29.254974	3979.826069	0.029867	74.314599	4117.290867	73.445335	112980.340539	5.496842e+08	51.899945	5.561032e+05
5	512.0	512.0	128.0	69.256516	8278.840379	0.071755	175.803029	8603.205054	173.864502	132259.555147	4.599156e+08	67.481513	4.309283e+05
6	512.0	1024.0	64.0	21.126702	4787.093086	0.016753	48.476630	4839.571319	43.521353	113433.281710	1.029340e+09	82.326924	1.021138e+06
7	512.0	1024.0	96.0	71.027394	5758.445127	0.059084	162.680963	5931.778747	146.218147	134612.067036	8.992812e+08	85.971068	8.986858e+05
8	512.0	1024.0	128.0	168.201720	10082.918573	0.143088	384.764061	10491.668301	344.232345	161279.290058	7.229306e+08	125.868647	7.224352e+05
9	729.0	256.0	64.0	3.919446	2117.714431	0.004471	10.619841	2153.829292	12.029603	89619.425074	3.720000e+08	36.863718	3.584134e+05
10	729.0	256.0	96.0	13.114453	3069.805048	0.014865	35.512586	3189.446788	36.792241	107839.480365	3.634820e+08	36.552259	3.746648e+05
11	729.0	256.0	128.0	31.020239	7356.841885	0.035716	83.967756	7639.153858	85.778220	119060.689391	3.171997e+08	58.581361	2.486982e+05
12	729.0	512.0	64.0	8.713507	3000.162207	0.008632	22.178372	3041.662247	22.819948	94113.593737	6.881474e+08	49.062660	6.398648e+05
13	729.0	512.0	96.0	29.255055	3959.201028	0.030058	74.390083	4096.648465	74.001554	107760.099818	4.966006e+08	54.511308	5.209588e+05
14	729.0	512.0	128.0	69.256597	8259.742252	0.071568	175.986288	8584.067250	170.873377	130377.047379	4.473117e+08	86.558061	4.347063e+05
15	729.0	1024.0	64.0	21.126783	4766.147990	0.016883	48.520073	4818.609902	45.182072	105824.610378	9.445332e+08	80.067664	9.269374e+05
16	729.0	1024.0	96.0	71.027475	5739.126709	0.060051	162.835093	5912.392497	148.806241	127325.061784	9.104156e+08	84.046033	9.057178e+05
17	729.0	1024.0	128.0	168.201801	10066.738985	0.143849	385.133740	10475.349482	344.477595	153577.675990	7.180802e+08	127.620555	7.121108e+05
18	1000.0	256.0	64.0	3.920205	2103.997804	0.004378	10.625269	2140.114192	12.208729	93100.388182	3.790373e+08	37.678360	3.875684e+05
19	1000.0	256.0	96.0	13.115211	3056.557325	0.015337	35.535707	3176.192967	38.435365	99740.296519	3.528847e+08	36.504454	2.697375e+05
20	1000.0	256.0	128.0	31.020998	7344.502784	0.036279	84.025333	7626.798315	87.961251	120896.697987	3.047861e+08	49.619193	3.042226e+05
21	1000.0	512.0	64.0	8.714266	2986.704635	0.008668	22.191250	3028.203227	24.303010	97909.120006	6.713427e+08	48.077980	6.148098e+05
22	1000.0	512.0	96.0	29.255814	3946.727065	0.030878	74.438348	4084.151843	76.918241	109081.277336	5.353949e+08	48.623166	5.490928e+05
23	1000.0	512.0	128.0	69.257356	8249.171712	0.072380	176.103465	8573.444772	173.974095	131365.308350	4.772688e+08	73.711210	4.433902e+05
24	1000.0	1024.0	64.0	21.127541	4753.276169	0.017586	48.547850	4805.717376	48.154416	108716.348388	1.023768e+09	76.088929	1.017975e+06
25	1000.0	1024.0	96.0	71.028233	5728.284651	0.060735	162.933646	5901.464608	149.567366	127912.972318	8.655995e+08	81.033593	8.637324e+05
26	1000.0	1024.0	128.0	168.202560	10059.813518	0.145373	385.370116	10468.247466	351.376095	154140.943753	6.691295e+08	114.951054	7.131300e+05

	Name	process	iteration	size	ex_directly	ex_indirectly	ex_real	in_directly	in_indirectly	in_real	vs directly(ex)	vs indirectly(ex)	vs directly(in)	vs indirectly(in)
0	.TAU_application	1000	1024	128	3.587083e-06	2.105709e-06	3.366667e-09	4.112130e+03	3.755078e+03	4.647539e+03	1.064470e+05	6.244581e+04	11.520269	1.920287e+01
1	MPI_Allreduce()	1000	1024	128	4.379160e+02	1.212476e+04	5.133010e-01	4.393460e+02	1.212476e+04	5.031910e+02	8.521368e+04	2.362014e+06	12.688031	2.309573e+03
2	MPI_Barrier()	1000	1024	128	8.163306e-04	1.850046e+01	3.833333e-05	8.691769e-04	1.850046e+01	2.200000e-02	2.029558e+03	4.826196e+07	96.049196	8.399298e+04
3	MPI_Comm_rank()	1000	1024	128	1.657494e-02	5.757170e-03	9.000000e-06	1.657544e-02	5.757170e-03	9.000000e-03	1.840660e+05	6.386856e+04	84.171572	3.603144e+01
4	MPI_Comm_size()	1000	1024	128	4.859529e-07	4.909538e-07	4.553333e-10	4.859529e-07	4.909538e-07	4.580000e-07	1.066247e+05	1.077229e+05	6.103262	7.195145e+00
5	MPI_Finalize()	1000	1024	128	4.884601e-01	1.307745e+05	7.433333e-04	5.010033e-01	1.307745e+05	6.880000e-01	6.561212e+04	1.759299e+10	27.179753	1.900783e+07
6	MPI_Init()	1000	1024	128	4.930194e-01	1.507272e+04	6.153333e-04	5.034056e-01	1.507272e+04	6.250000e-01	8.002234e+04	2.449521e+09	19.455111	2.411535e+06
7	MPI_Irecv()	1000	1024	128	6.071899e-02	-8.960680e-02	5.633333e-05	6.355979e-02	-8.960680e-02	7.500000e-02	1.076852e+05	1.591653e+05	15.253612	2.194757e+02
8	MPI_Isend()	1000	1024	128	3.174987e+00	1.111102e+00	8.950000e-04	4.263862e-01	1.111102e+00	8.900000e-01	3.546471e+05	1.240455e+05	52.091433	2.484295e+01
9	MPI_Reduce()	1000	1024	128	5.818206e-02	7.866580e+02	7.166667e-05	6.009330e-02	7.866580e+02	7.100000e-02	8.108427e+04	1.097662e+09	15.361552	1.107869e+06
10	MPI_Wait()	1000	1024	128	7.150075e+01	1.164307e+01	4.289367e-02	7.204855e+01	1.164307e+01	4.386800e+01	1.665930e+05	2.704403e+04	64.239424	7.345886e+01
11	MPI_Waitall()	1000	1024	128	1.919569e+03	1.592230e+05	1.180570e-01	2.087961e+03	1.592230e+05	1.217270e+02	1.625868e+06	1.348695e+08	1615.281445	1.307034e+05
12	Real_t_CalcElemVolume(const	1000	1024	128	3.353785e+02	2.937287e+02	3.164563e-01	3.370815e+02	2.937287e+02	3.166190e+02	1.058794e+05	9.271810e+04	6.462806	7.229590e+00
13	StrToInt	1000	1024	128	1.918167e-06	2.178581e-06	2.020000e-09	1.918167e-06	2.178581e-06	2.000000e-06	9.485875e+04	1.077505e+05	4.091667	8.929028e+00
14	int_main(int_char_**)	1000	1024	128	2.135894e+03	2.103485e+03	3.181085e+00	4.112131e+03	3.755078e+03	4.647539e+03	6.704358e+04	6.602477e+04	11.520242	1.920287e+01
15	void_CalcKinematicsForElems(Domain	1000	1024	128	4.454195e+02	6.681816e+02	4.528560e-01	7.486615e+02	1.138635e+03	7.694730e+02	9.825788e+04	1.474484e+05	2.704650	4.797598e+01
16	void_CommMonoQ(Domain	1000	1024	128	5.258543e+00	2.516016e+01	1.440000e-03	5.863060e+01	3.765066e+00	5.362000e+00	3.650766e+05	1.747134e+06	993.446434	2.978243e+01
17	void_CommRecv(Domain	1000	1024	128	1.955217e-03	-3.676996e-01	4.700000e-05	9.793986e-02	9.577802e-02	9.500000e-02	4.060035e+03	7.824395e+05	3.094587	8.189667e-01
18	void_CommSBN(Domain	1000	1024	128	1.572116e+00	1.703929e+00	2.805667e-03	5.900257e-01	2.565895e-01	2.400900e+01	5.593361e+04	6.063171e+04	97.542481	9.893128e+01
19	void_CommSend(Domain	1000	1024	128	3.134054e+00	4.321765e+02	8.500000e-03	3.198355e+02	7.305781e+03	1.310700e+02	3.677122e+04	5.084330e+06	144.018843	5.473953e+03
20	void_CommSyncPosVel(Domain	1000	1024	128	2.176785e+01	9.083651e+01	1.143733e-02	4.040905e+01	2.745875e+02	3.012900e+01	1.902228e+05	7.941106e+05	34.120124	8.113728e+02
21	void_Domain::BuildMesh(Int_t_Int_t_Int_t)	1000	1024	128	1.012948e-02	3.854139e-01	5.866667e-05	1.025539e-02	3.854139e-01	4.600000e-02	1.716616e+04	6.568556e+05	77.705669	7.378564e+02
22	void_Domain::CreateRegionIndexSets(Int_t_Int_t)	1000	1024	128	7.220721e-03	5.744107e-02	2.166667e-05	7.634154e-03	5.741094e-02	2.100000e-02	3.322640e+04	2.650126e+05	63.646887	1.733854e+02
23	void_Domain::Domain(Int_t_Index_t_Index_t_Inde...	1000	1024	128	7.318609e-01	2.834858e+02	5.293333e-04	1.303911e+00	4.412710e+02	9.360000e-01	1.381609e+05	5.355515e+07	39.306766	4.704434e+04
24	void_Domain::SetupBoundaryConditions(Int_t)	1000	1024	128	4.859352e-03	2.338127e-02	7.000000e-06	4.859352e-03	2.338127e-02	7.000000e-03	6.931931e+04	3.339181e+05	30.580687	2.340181e+02
25	void_Domain::SetupCommBuffers(Int_t)	1000	1024	128	4.862345e-03	9.720498e-02	4.000000e-06	5.040010e-03	9.720498e-02	4.000000e-03	1.214586e+05	2.430025e+06	26.000254	2.330125e+03
26	void_Domain::SetupElementConnectivities(Int_t)	1000	1024	128	1.719932e-02	2.235052e+00	1.866667e-05	1.716865e-02	2.235052e+00	1.900000e-02	9.203924e+04	1.197339e+07	9.638680	1.166343e+04
27	void_Domain::SetupSymmetryPlanes(Int_t)	1000	1024	128	3.860905e-05	3.889435e-05	2.836667e-08	3.883078e-05	3.889435e-05	2.820000e-05	1.360071e+05	1.370129e+05	37.697803	3.792323e+01
28	void_Domain::~Domain()	1000	1024	128	4.324454e-04	7.906917e-04	1.000000e-06	4.324454e-04	7.906917e-04	1.000000e-03	4.314454e+04	7.896917e+04	56.755456	2.093083e+01
29	void_InitMeshDecomp(Int_t_Int_t_Int_t	1000	1024	128	1.127706e-05	1.083020e-05	1.226667e-08	1.127706e-05	1.083020e-05	1.220000e-05	9.183252e+04	8.818964e+04	7.565118	1.122790e+01
30	void_ParseCommandLineOptions(int_char_**_Int_t...	1000	1024	128	2.085278e-06	2.016489e-06	1.573333e-09	3.947028e-06	4.734345e-06	3.560000e-06	1.324388e+05	1.280666e+05	10.871567	3.298721e+01
31	void_VerifyAndWriteFinalOutput(Real_t_Domain	1000	1024	128	9.571972e-08	9.605993e-06	1.296667e-10	1.368243e-07	9.605993e-06	1.400000e-07	7.371984e+04	7.408121e+06	2.268338	6.761423e+03

	all to all	perCall to all	real inclusive	real call	vs all to all	vs perCall to all	process	iteration	size	weighted all	weighted perCall
0	10.611352	9.676474	10.839249	2.106369e+06	37.292687	35.430187	512.0	256.0	64.0	7.737990	11.677585
1	35.476425	31.471555	36.179501	7.106557e+06	38.620858	38.064345	512.0	256.0	96.0	12.185571	16.879060
2	83.877708	73.667464	85.971731	1.684378e+07	54.544020	45.268959	512.0	256.0	128.0	15.372071	20.704700
3	22.158231	19.175062	21.592166	4.204548e+06	49.470312	44.051874	512.0	512.0	64.0	12.806895	12.086386
4	74.314599	62.740502	73.697679	1.418549e+07	54.582365	42.422741	512.0	512.0	96.0	18.343714	18.276245
5	175.803029	147.083379	172.934481	3.362205e+07	73.606628	45.303635	512.0	512.0	128.0	20.484699	21.012191
6	48.476630	38.176309	44.166937	8.400906e+06	80.896379	63.477555	512.0	1024.0	64.0	24.250173	14.151128
7	162.680963	125.282475	145.128272	2.834328e+07	89.564398	51.254591	512.0	1024.0	96.0	27.234583	17.369565
8	384.764061	293.919305	344.319085	6.717847e+07	127.674076	53.889614	512.0	1024.0	128.0	30.057984	20.883720
9	10.619841	9.675814	12.003718	2.106394e+06	36.404214	37.327517	729.0	256.0	64.0	15.376221	20.442626
10	35.512586	31.470908	36.524022	7.106582e+06	38.626855	39.609083	729.0	256.0	96.0	12.594487	17.364630
11	83.967756	73.666845	85.826887	1.684380e+07	60.144997	47.124122	729.0	256.0	128.0	15.207501	20.652246
12	22.178372	19.174402	22.966625	4.204598e+06	48.673313	39.509351	729.0	512.0	64.0	16.780932	17.162932
13	74.390083	62.739855	74.222449	1.418554e+07	53.314672	41.362377	729.0	512.0	96.0	18.602162	18.647579
14	175.986288	147.082760	172.148075	3.362210e+07	79.305981	48.371853	729.0	512.0	128.0	20.222788	20.873158
15	48.520073	38.175649	44.713030	8.401006e+06	79.894371	59.372170	729.0	1024.0	64.0	25.170952	16.156161
16	162.835093	125.281828	148.186731	2.834338e+07	83.965075	52.108529	729.0	1024.0	96.0	28.213289	18.685817
17	385.133740	293.918686	347.009533	6.717857e+07	118.948031	53.259494	729.0	1024.0	128.0	30.140738	21.168359
18	10.625269	9.675391	11.676583	2.106414e+06	37.717062	37.350154	1000.0	256.0	64.0	13.037556	18.212172
19	35.535707	31.470494	38.129689	7.106602e+06	34.052828	34.529437	1000.0	256.0	96.0	15.273942	20.061783
20	84.025333	73.666448	87.698720	1.684382e+07	51.917553	43.952556	1000.0	256.0	128.0	16.429116	21.834959
21	22.191250	19.173979	23.075291	4.204639e+06	47.847240	39.671418	1000.0	512.0	64.0	17.187382	18.132954
22	74.438348	62.739441	77.022960	1.418558e+07	49.269650	41.011409	1000.0	512.0	96.0	20.790522	20.926576
23	176.103465	147.082363	174.806887	3.362214e+07	73.577840	45.496515	1000.0	512.0	128.0	21.001375	21.694538
24	48.547850	38.175226	46.963656	8.401087e+06	77.625400	60.831823	1000.0	1024.0	64.0	27.506611	19.248749
25	162.933646	125.281414	150.549970	2.834346e+07	80.981623	49.609760	1000.0	1024.0	96.0	28.812038	19.511240
26	385.370116	293.918289	351.664731	6.717865e+07	117.023694	52.677614	1000.0	1024.0	128.0	30.867999	22.073089

	Name	process	iteration	size	_all	_perCall	_call	_real_call	_real_in	vs _all	vs _perCall
0	.TAU_application	8	8	16	8.33841224987527	2.26833470751862	0.999999999999999	1.000000	6.966333e+00	19.6958550630451	67.4386137013453
1	MPI_Allreduce()	8	8	16	0.0351393561255279	0.0129322982170460	6.99999999999999	7.000000	5.000000e-03	602.787122510559	158.645964340919
2	MPI_Barrier()	8	8	16	0.000869176944444444	0.000775535421489941	0.999999999999999	1.000000	5.146667e-05	1588.81530656304	1406.86934227320
3	MPI_Comm_rank()	8	8	16	2.32389784946236e-5	1.91007170619486e-5	77.0000000000001	77.000000	1.700000e-05	36.6998734977860	12.3571591879332
4	MPI_Comm_size()	8	8	16	4.51732307973519e-7	3.51328287732536e-7	0.999999999999999	1.000000	2.500000e-07	80.6929231894076	40.5313150930144
...	...	...	...	...	...	...	...	...	...	...	...
3835	void_Domain::SetupSymmetryPlanes(Int_t)	343	128	48	5.75602235343737e-6	5.71681779839729e-6	0.999999999999999	1.000000	5.483333e-06	4.97305203837142	4.25807535071048
3836	void_Domain::~Domain()	343	128	48	0.000127211158730159	0.000125590400000000	0.999999999999999	1.000000	1.740000e-04	26.8901386608284	27.8216091954024
3837	void_InitMeshDecomp(Int_t_Int_t_Int_t	343	128	48	1.12770555555556e-5	1.12723333333333e-5	0.999999999999999	1.000000	1.160000e-05	2.78400383141756	2.82471264367818
3838	void_ParseCommandLineOptions(int_char_**_Int_t...	343	128	48	3.94702777777778e-6	4.02879424283154e-6	0.999999999999999	1.000000	3.506667e-06	12.5578263624841	14.8895696624963
3839	void_VerifyAndWriteFinalOutput(Real_t_Domain	343	128	48	2.99538480025158e-7	2.56034860759274e-7	0.00291544582073221	0.002915	2.703333e-07	10.8033834864950	5.28920070557076

	all to all	perCall to all	real inclusive	real call	vs all to all	vs perCall to all	Name
0	6.499275e+00	5.996800e+00	6.499275e+00	1.000000e+00	18.531208	47.868233	.TAU_application
1	4.851583e-01	4.998763e-01	4.851583e-01	4.860000e+01	70.028502	54.303651	MPI_Allreduce()
2	8.691769e-04	8.869008e-04	8.691769e-04	1.000000e+00	396.747887	409.610224	MPI_Barrier()
3	1.959781e-04	2.089635e-04	1.959781e-04	4.514000e+02	13.912200	21.113674	MPI_Comm_rank()
4	4.641972e-07	4.599250e-07	4.641972e-07	1.000000e+00	14.603987	14.554590	MPI_Comm_size()
5	2.350408e+00	1.864908e+00	2.350408e+00	1.000000e+00	57.047656	72.380782	MPI_Finalize()
6	5.034056e-01	4.953750e-01	5.034056e-01	1.000000e+00	19.225513	18.808969	MPI_Init()
7	8.598322e-04	8.888162e-04	8.598322e-04	1.310194e+03	30.815303	33.889409	MPI_Irecv()
8	1.004190e-02	8.991201e-03	1.004190e-02	1.310194e+03	196.520935	168.800569	MPI_Isend()
9	1.340376e-03	1.195829e-03	1.340376e-03	1.000000e+00	711.239913	608.350769	MPI_Reduce()
10	8.030106e-02	7.559656e-02	8.030106e-02	1.310194e+03	670.223115	593.371131	MPI_Wait()
11	2.520304e-01	2.651487e-01	2.520304e-01	1.498000e+02	235.469099	227.909330	MPI_Waitall()
12	3.230556e-01	3.055593e-01	3.230556e-01	2.040194e+06	26.653478	9.741670	Real_t_CalcElemVolume(const
13	1.918167e-06	1.896250e-06	1.918167e-06	2.000000e+00	4.813321	4.558068	StrToInt
14	6.499272e+00	5.996800e+00	6.499272e+00	1.000000e+00	18.531235	47.868432	int_main(int_char_**)
15	7.321944e-01	7.243828e-01	7.321944e-01	4.960000e+01	20.147096	8.701478	void_CalcKinematicsForElems(Domain
16	1.996276e-02	2.119956e-02	1.996276e-02	4.960000e+01	337.411414	587.220889	void_CommMonoQ(Domain
17	1.326047e-03	1.262148e-03	1.326047e-03	1.498000e+02	28.323250	27.543479	void_CommRecv(Domain
18	3.238175e-02	3.545512e-02	3.238175e-02	5.060000e+01	653.772904	647.333088	void_CommSBN(Domain
19	2.972667e-01	2.995477e-01	2.972667e-01	1.498000e+02	175.679522	174.153456	void_CommSend(Domain
20	5.111839e-02	4.868275e-02	5.111839e-02	4.960000e+01	187.353032	142.852551	void_CommSyncPosVel(Domain
21	5.212414e-04	4.756567e-04	5.212414e-04	1.000000e+00	26.816748	30.119448	void_Domain::BuildMesh(Int_t_Int_t_Int_t)
22	4.200039e-04	4.166692e-04	4.200039e-04	1.000000e+00	7.963392	7.352371	void_Domain::CreateRegionIndexSets(Int_t_Int_t)
23	2.445430e-02	2.351667e-02	2.445430e-02	1.000000e+00	16.558680	12.883915	void_Domain::Domain(Int_t_Index_t_Index_t_Inde...
24	8.139061e-05	7.854017e-05	8.139061e-05	1.000000e+00	61.170717	56.710331	void_Domain::SetupBoundaryConditions(Int_t)
25	3.473678e-04	3.449475e-04	3.473678e-04	1.000000e+00	22.790014	22.831559	void_Domain::SetupCommBuffers(Int_t)
26	3.219914e-04	3.211900e-04	3.219914e-04	1.000000e+00	5.643632	5.133163	void_Domain::SetupElementConnectivities(Int_t)
27	2.786811e-06	2.776750e-06	2.786811e-06	1.000000e+00	6.538504	6.346429	void_Domain::SetupSymmetryPlanes(Int_t)
28	5.853344e-05	5.811650e-05	5.853344e-05	1.000000e+00	587.887615	591.632391	void_Domain::~Domain()
29	1.127706e-05	1.127233e-05	1.127706e-05	1.000000e+00	5.037710	5.035600	void_InitMeshDecomp(Int_t_Int_t_Int_t
30	3.947028e-06	3.958000e-06	3.947028e-06	1.000000e+00	12.453269	12.553923	void_ParseCommandLineOptions(int_char_**_Int_t...
31	2.787308e-06	2.827912e-06	2.787308e-06	3.220118e-02	9.054462	5.655438	void_VerifyAndWriteFinalOutput(Real_t_Domain

	all to all	perCall to all	real exclusive	real call	vs all to all	vs perCall to all	process	iteration	size	weighted all	weighted perCall
0	3.919365	3.562725	4.129666	2.106369e+06	33.035133	35.831271	512.0	256.0	64.0	20.742996	19.959976
1	13.114372	11.462402	14.749491	7.106557e+06	43.461620	37.633187	512.0	256.0	96.0	27.631758	27.335283
2	31.020158	26.597047	35.743929	1.684378e+07	61.324288	44.530961	512.0	256.0	128.0	33.291483	33.645807
3	8.713426	7.086378	8.229333	4.204548e+06	35.542758	56.707942	512.0	512.0	64.0	32.149903	19.966094
4	29.254974	22.871801	29.867408	1.418549e+07	50.781096	42.295487	512.0	512.0	96.0	38.626849	27.925843
5	69.256516	53.112093	71.755460	3.362205e+07	72.430286	59.479508	512.0	512.0	128.0	43.509452	33.603933
6	21.126702	14.139320	16.752916	8.400906e+06	52.645392	107.862594	512.0	1024.0	64.0	55.409869	21.430934
7	71.027394	45.701645	59.083991	2.834328e+07	68.856654	107.176906	512.0	1024.0	96.0	59.680412	27.542671
8	168.201720	106.161514	143.088033	6.717847e+07	102.900533	117.750802	512.0	1024.0	128.0	64.545745	33.720302
9	3.919446	3.562086	4.470895	2.106394e+06	28.337839	33.316542	729.0	256.0	64.0	19.142831	21.792942
10	13.114453	11.461814	14.864783	7.106582e+06	46.335564	40.238899	729.0	256.0	96.0	27.847072	27.522685
11	31.020239	26.596560	35.715866	1.684380e+07	63.040293	43.229220	729.0	256.0	128.0	33.352645	33.778572
12	8.713507	7.085757	8.631677	4.204598e+06	32.987539	46.299854	729.0	512.0	64.0	29.016446	19.862770
13	29.255055	22.871269	30.057960	1.418554e+07	50.990390	42.400752	729.0	512.0	96.0	38.743319	28.098067
14	69.256597	53.111736	71.567647	3.362210e+07	76.383717	53.270361	729.0	512.0	128.0	43.697301	33.829149
15	21.126783	14.138728	16.882812	8.401006e+06	48.957995	101.215953	729.0	1024.0	64.0	50.151774	18.097539
16	71.027475	45.701218	60.051387	2.834338e+07	68.194426	108.782021	729.0	1024.0	96.0	59.832136	28.208945
17	168.201801	106.161410	143.849095	6.717857e+07	99.479951	73.517453	729.0	1024.0	128.0	64.301007	33.695873
18	3.920205	3.561676	4.377718	2.106414e+06	28.275866	33.864247	1000.0	256.0	64.0	17.709084	20.275724
19	13.115211	11.461438	15.337157	7.106602e+06	40.035304	38.153890	1000.0	256.0	96.0	28.767566	28.708202
20	31.020998	26.596248	36.279199	1.684382e+07	61.212865	43.461281	1000.0	256.0	128.0	33.728206	34.070148
21	8.714266	7.085359	8.668156	4.204639e+06	36.364071	47.253969	1000.0	512.0	64.0	32.809628	19.992638
22	29.255814	22.870929	30.877772	1.418558e+07	49.038449	45.953220	1000.0	512.0	96.0	39.506253	29.168207
23	69.257356	53.111507	72.379939	3.362214e+07	71.829526	48.820699	1000.0	512.0	128.0	43.823675	34.006411
24	21.127541	14.138348	17.585510	8.401087e+06	45.927345	100.938777	1000.0	1024.0	64.0	50.567381	21.253765
25	71.028233	45.700944	60.735022	2.834346e+07	66.114713	105.331270	1000.0	1024.0	96.0	59.713227	28.463254
26	168.202560	106.161342	145.373387	6.717865e+07	98.420003	76.186773	1000.0	1024.0	128.0	64.507110	34.165177

	Name	process	iteration	size	_all	_perCall	_real_call	_call	_real_ex	vs _all	vs _perCall
0	.TAU_application	8	8	16	3.58708333333333e-6	3.68817545339519e-6	1.000000	0.999999999999999	3.420000e-06	4.88547758284604	7.84138752617504
1	MPI_Allreduce()	8	8	16	0.0230764914454476	0.0129322982170460	7.000000	6.99999999999999	5.000000e-03	361.529828908953	158.645964340919
2	MPI_Barrier()	8	8	16	0.000816330555555555	0.000775535421489941	1.000000	0.999999999999999	5.140000e-05	1488.19174232598	1408.82377721778
3	MPI_Comm_rank()	8	8	16	2.32349761051373e-5	1.91007170619486e-5	77.000000	77.0000000000001	1.700000e-05	36.6763300302195	12.3571591879332
4	MPI_Comm_size()	8	8	16	4.51732307973519e-7	3.51328287732536e-7	1.000000	0.999999999999999	2.500000e-07	80.6929231894076	40.5313150930144
...	...	...	...	...	...	...	...	...	...	...	...
3835	void_Domain::SetupSymmetryPlanes(Int_t)	343	128	48	5.72710187918834e-6	5.71681779839729e-6	1.000000	0.999999999999999	5.483333e-06	4.44562697607912	4.25807535071048
3836	void_Domain::~Domain()	343	128	48	0.000127211158730159	0.000125590400000000	1.000000	0.999999999999999	1.740000e-04	26.8901386608284	27.8216091954024
3837	void_InitMeshDecomp(Int_t_Int_t_Int_t	343	128	48	1.12770555555556e-5	1.12723333333333e-5	1.000000	0.999999999999999	1.160000e-05	2.78400383141756	2.82471264367818
3838	void_ParseCommandLineOptions(int_char_**_Int_t...	343	128	48	2.08527777777778e-6	2.06058333333333e-6	1.000000	0.999999999999999	1.606667e-06	29.7890733056709	28.2520746887966
3839	void_VerifyAndWriteFinalOutput(Real_t_Domain	343	128	48	2.58210082572280e-7	2.56034860759274e-7	0.002915	0.00291544582073221	2.703333e-07	4.48455638509998	5.28920070557076

	all to all	perCall to all	real exclusive	real call	vs all to all	vs perCall to all	Name
0	3.587083e-06	3.611250e-06	3.587083e-06	1.000000e+00	4.494972	4.529864	.TAU_application
1	4.656926e-01	4.998763e-01	4.656926e-01	4.860000e+01	64.186959	61.089047	MPI_Allreduce()
2	8.163306e-04	8.869008e-04	8.163306e-04	1.000000e+00	375.845370	413.759584	MPI_Barrier()
3	1.959725e-04	2.089635e-04	1.959725e-04	4.514000e+02	13.913296	21.117243	MPI_Comm_rank()
4	4.641972e-07	4.599250e-07	4.641972e-07	1.000000e+00	14.603987	14.554590	MPI_Comm_size()
5	2.341597e+00	1.864908e+00	2.341597e+00	1.000000e+00	58.939041	76.746281	MPI_Finalize()
6	4.930194e-01	4.953750e-01	4.930194e-01	1.000000e+00	22.840952	22.984320	MPI_Init()
7	8.281572e-04	8.888162e-04	8.281572e-04	1.310194e+03	29.905130	34.624772	MPI_Irecv()
8	8.510761e-03	8.991201e-03	8.510761e-03	1.310194e+03	194.124650	173.789329	MPI_Isend()
9	1.277368e-03	1.195829e-03	1.277368e-03	1.000000e+00	680.716426	609.283209	MPI_Reduce()
10	7.208439e-02	7.559656e-02	7.208439e-02	1.310194e+03	637.536923	676.371267	MPI_Wait()
11	2.359443e-01	2.651487e-01	2.359443e-01	1.498000e+02	238.412718	229.118919	MPI_Waitall()
12	3.101924e-01	3.055593e-01	3.101924e-01	2.040194e+06	16.134983	15.743301	Real_t_CalcElemVolume(const
13	1.918167e-06	1.896250e-06	1.918167e-06	2.000000e+00	4.813321	4.558068	StrToInt
14	1.996847e+00	1.984633e+00	1.996847e+00	1.000000e+00	7.036762	13.104779	int_main(int_char_**)
15	4.425535e-01	4.260046e-01	4.425535e-01	4.960000e+01	63.548398	16.200636	void_CalcKinematicsForElems(Domain
16	5.124423e-03	3.397645e-03	5.124423e-03	4.960000e+01	55.767147	37.614828	void_CommMonoQ(Domain
17	5.466008e-04	3.667988e-04	5.466008e-04	1.498000e+02	189.180976	22.339178	void_CommRecv(Domain
18	9.188817e-03	3.892476e-03	9.188817e-03	5.060000e+01	411.382561	39.753054	void_CommSBN(Domain
19	7.149106e-02	1.618499e-02	7.149106e-02	1.498000e+02	847.548282	54.300345	void_CommSend(Domain
20	2.103545e-02	1.281131e-02	2.103545e-02	4.960000e+01	147.489721	83.518617	void_CommSyncPosVel(Domain
21	5.171553e-04	4.756567e-04	5.171553e-04	1.000000e+00	25.108523	29.898369	void_Domain::BuildMesh(Int_t_Int_t_Int_t)
22	4.043950e-04	4.160833e-04	4.043950e-04	1.000000e+00	9.228406	9.643183	void_Domain::CreateRegionIndexSets(Int_t_Int_t)
23	1.408947e-02	1.435833e-02	1.408948e-02	1.000000e+00	29.694414	23.957419	void_Domain::Domain(Int_t_Index_t_Index_t_Inde...
24	8.139061e-05	7.854017e-05	8.139061e-05	1.000000e+00	61.170717	56.710331	void_Domain::SetupBoundaryConditions(Int_t)
25	3.404347e-04	3.449475e-04	3.404347e-04	1.000000e+00	25.141755	25.130508	void_Domain::SetupCommBuffers(Int_t)
26	3.176192e-04	3.211900e-04	3.176192e-04	1.000000e+00	4.971856	5.742696	void_Domain::SetupElementConnectivities(Int_t)
27	2.775200e-06	2.776750e-06	2.775200e-06	1.000000e+00	6.611437	6.463983	void_Domain::SetupSymmetryPlanes(Int_t)
28	5.853344e-05	5.811650e-05	5.853344e-05	1.000000e+00	587.887615	591.632391	void_Domain::~Domain()
29	1.127706e-05	1.127233e-05	1.127706e-05	1.000000e+00	5.037710	5.035600	void_InitMeshDecomp(Int_t_Int_t_Int_t
30	2.085278e-06	2.060583e-06	2.085278e-06	1.000000e+00	25.389755	24.495431	void_ParseCommandLineOptions(int_char_**_Int_t...
31	2.742558e-06	2.827912e-06	2.742558e-06	3.220118e-02	6.014462	6.400693	void_VerifyAndWriteFinalOutput(Real_t_Domain

	予測値	実測値	MAPE	関数名
0	1	1	1.11e-13	.TAU_application
1	48.6	48.6	4.056e-14	MPI_Allreduce()
2	1	1	1.11e-13	MPI_Barrier()
3	451.4	451.4	6.455e-14	MPI_Comm_rank()
4	1	1	1.11e-13	MPI_Comm_size()
5	1	1	1.11e-13	MPI_Finalize()
6	1	1	1.11e-13	MPI_Init()
7	1310	1310	28.14	MPI_Irecv()
8	1310	1310	28.14	MPI_Isend()
9	1	1	1.11e-13	MPI_Reduce()
10	1310	1310	28.14	MPI_Wait()
11	149.8	149.8	3.22e-14	MPI_Waitall()
12	2.04e+06	2.04e+06	0.0004923	Real_t_CalcElemVolume(const
13	2	2	1.11e-13	StrToInt
14	1	1	1.11e-13	int_main(int_char_**)
15	49.6	49.6	0	void_CalcKinematicsForElems(Domain
16	49.6	49.6	0	void_CommMonoQ(Domain
17	149.8	149.8	3.22e-14	void_CommRecv(Domain
18	50.6	50.6	1.703e-14	void_CommSBN(Domain
19	149.8	149.8	3.22e-14	void_CommSend(Domain
20	49.6	49.6	0	void_CommSyncPosVel(Domain
21	1	1	1.11e-13	void_Domain::BuildMesh(Int_t_Int_t_Int_t)
22	1	1	1.11e-13	void_Domain::CreateRegionIndexSets(Int_t_Int_t)
23	1	1	1.11e-13	void_Domain::Domain(Int_t_Index_t_Index_t_Inde...
24	1	1	1.11e-13	void_Domain::SetupBoundaryConditions(Int_t)
25	1	1	1.11e-13	void_Domain::SetupCommBuffers(Int_t)
26	1	1	1.11e-13	void_Domain::SetupElementConnectivities(Int_t)
27	1	1	1.11e-13	void_Domain::SetupSymmetryPlanes(Int_t)
28	1	1	1.11e-13	void_Domain::~Domain()
29	1	1	1.11e-13	void_InitMeshDecomp(Int_t_Int_t_Int_t
30	1	1	1.11e-13	void_ParseCommandLineOptions(int_char_**_Int_t...
31	0.0322	0.0322	8.135e-05	void_VerifyAndWriteFinalOutput(Real_t_Domain

	real_Exclusive	real_call	predicted_Exclusive_all	predicted_Exclusive_perCall	predicted_call	process	problem_size	functionName
0	2.420000e-05	1.000000	8.11392857142857e-6	8.11392857142857e-6	1.00000000000000	256	128	.TAU_application
1	2.820000e-04	1.000000	0.000242821428571429	0.000242821428571428	1.00000000000000	256	128	MG_MPI
2	5.380000e-01	1.000000	0.365928571428571	0.365928571428571	1.00000000000000	256	128	MPI_Init()
3	5.220000e-01	1.000000	0.395925592813830	0.395925592813830	1.00000000000000	256	128	MPI_Finalize()
4	5.730000e-04	51.000000	0.000432857142857143	0.000445989795918367	51.0000000000000	256	128	MG3P
...	...	...	...	...	...	...	...	...
118	1.600000e-06	2.000000	9.70451658521133e-7	9.70451658521133e-7	2.00000000000000	256	512	TIMER_STOP
119	5.510000e-07	0.003906	-1.16555286677527e-6	2.66765644798171e-7	0.00390624999999999	256	512	PRINT_RESULTS
120	4.530000e-07	1.000000	3.44357142857143e-7	3.44357142857143e-7	1.00000000000000	256	512	MPI_Comm_size()
121	2.540000e-07	1.000000	2.20428571428571e-7	2.20428571428571e-7	1.00000000000000	256	512	MPI_Comm_rank()
122	1.520000e-07	1.003910	2.27142857142857e-7	2.01702807319246e-7	1.00390448796271	256	512	TIMER_READ

	real_Exclusive	real_call	predicted_Exclusive_all	predicted_Exclusive_perCall	predicted_call	process	problem_size	functionName	vs all	vs perCall
0	2.420000e-05	1.000000	8.11392857142857e-6	8.11392857142857e-6	1.00000000000000	256	128	.TAU_application	66.4713695395514	66.4713695395514
1	2.820000e-04	1.000000	0.000242821428571429	0.000242821428571428	1.00000000000000	256	128	MG_MPI	13.8931104356636	13.8931104356637
2	5.380000e-01	1.000000	0.365928571428571	0.365928571428571	1.00000000000000	256	128	MPI_Init()	31.9835369091875	31.9835369091875
3	5.220000e-01	1.000000	0.395925592813830	0.395925592813830	1.00000000000000	256	128	MPI_Finalize()	24.1521852847068	24.1521852847069
4	5.730000e-04	51.000000	0.000432857142857143	0.000445989795918367	51.0000000000000	256	128	MG3P	24.4577412116679	22.1658296826584
...	...	...	...	...	...	...	...	...	...	...
118	1.600000e-06	2.000000	9.70451658521133e-7	9.70451658521133e-7	2.00000000000000	256	512	TIMER_STOP	39.3467713424292	39.3467713424292
119	5.510000e-07	0.003906	-1.16555286677527e-6	2.66765644798171e-7	0.00390624999999999	256	512	PRINT_RESULTS	311.534095603497	51.5851824322738
120	4.530000e-07	1.000000	3.44357142857143e-7	3.44357142857143e-7	1.00000000000000	256	512	MPI_Comm_size()	23.9829706717124	23.9829706717124
121	2.540000e-07	1.000000	2.20428571428571e-7	2.20428571428571e-7	1.00000000000000	256	512	MPI_Comm_rank()	13.2170978627671	13.2170978627672
122	1.520000e-07	1.003910	2.27142857142857e-7	2.01702807319246e-7	1.00390448796271	256	512	TIMER_READ	49.4360902255639	32.6992153416093